How to Run Call of Cthulhu 7e: Keeper Guide to Sessions, NPCs, and Campaign Pacing

Introduction

Running Call of Cthulhu 7e means preparing a mystery with enough structure that player choices produce consequences, then presenting it clearly enough that investigators can understand what is happening before the danger consumes them. You are not there to protect a plot from the players.

The Keeper has more moving parts than any player: the scenario's truth, the clues, the threats, the people with competing motives, the rules, the tone, and the pace of the session. Exhaustive preparation will not cover all of that. Prepare the things that let you respond well: what happened, what changes if no one intervenes, what evidence can reveal it, and which people will make the investigators care.

This guide covers the whole job in one place: preparing a session, running it minute to minute, building NPCs, and pacing a campaign across many cases. Rules procedures are checked against the Keeper Rulebook. Everything else is tablecraft you can adapt.

Hot Take

The Keeper's job is to make success costly, intelligible, and worth pursuing. Making the investigators fail is not the job.

A dead end is not horror. A clue that requires the players to guess the Keeper's exact thought is not mystery. A monster that kills the group before anyone can react is not tension. Call of Cthulhu becomes frightening when players see enough of the threat to make a choice, understand that every available choice has a price, and act anyway because someone or something needs them to.

Prepare pressure, not punishment. The best sessions leave players saying, "We should not have opened that door," while also knowing why they opened it.

What the Keeper Is Responsible For

The Keeper presents the world, portrays everyone in it who is not an investigator, adjudicates rules, and knows the underlying situation. Think of yourself as the moderator of a shared horror story. You know the hidden truth and play the world around it, and that authority exists to produce fairness and responsiveness.

The practical rule is this: be honest about what the investigators can perceive, clear about what their choices risk, and flexible about how they pursue the truth. You can keep secrets inside the fiction. You should not make the players guess basic environmental facts, the likely consequence of an obvious action, or the only route you will accept as progress.

This matters more in Call of Cthulhu than in many games because investigators are not durable heroes. A bad decision may cost Sanity, safety, reputation, a relationship, or a life. That cost makes excellent horror when it follows from information the group had. It becomes frustrating when it arrives as a punishment for not reading the Keeper's mind.

The Keeper Rulebook backs this up with its dice rules: the Keeper decides when a roll is needed and how hard it is, a routine task with no drama needs no roll, and losing a roll does not automatically mean failing the goal. The Keeper decides what a lost roll means, and the book warns against declaring outcomes that block play.

Before a session, know five things:

1. The truth. What happened before the investigators arrived? Who caused it, what do they want, and what are they hiding?
2. The active threat. What will happen next if the investigators do nothing? Give it a timeline, an escalation, or a decision point.
3. The evidence. Which physical facts, witnesses, records, locations, and consequences can reveal the truth?
4. The people. Who wants help, who wants silence, who is afraid, and who profits from confusion?
5. The likely costs. What risks follow investigation: time, exposure, injury, legal trouble, Sanity, relationships, or an enemy becoming aware of them?

You do not need to know every scene. You need to know enough that a new scene follows logically from an investigator's action. A player who visits an unplanned location should not meet a blank wall simply because it was not on your map. Ask what the location reasonably contains, what the antagonists would have done, and what existing evidence could point there.

Start With a Table Agreement

Horror only works when the people at the table trust each other. Before the campaign or one-shot, establish the expected tone, the era, the type of investigation, and the content boundaries. This tells you what kind of fear is welcome and what kind will make the session worse for someone.

Discuss at least these subjects:

• The intended tone: slow-burn dread, lurid pulp, mystery with occasional horror, tragic drama, or a mix.
• Topics to exclude completely and topics to fade to black when they appear.
• How closely the game will depict historical prejudice, violence, medical harm, children in danger, cult coercion, and other sensitive material.
• Whether player-versus-player conflict, betrayal, or secret objectives are welcome.
• What tools the group will use to pause, revise, or step away from a scene.
• Whether the investigators begin as a connected group or meet through the scenario.

Historical settings deserve particular care. The 1920s contained real discrimination and systems of exclusion. You may acknowledge that history, abstract it, or choose an alternate presentation, but make the decision with the group rather than surprising people with it during a supposedly recreational evening. No amount of period accuracy makes a table boundary optional.

Keep the physical or virtual table functional too. Have character sheets, dice, a way to show handouts, and lighting that lets everyone read. Music, props, and dim lights can support tone, but they are optional and must never make communication or safety tools harder to use. A clear table beats an elaborate atmosphere that leaves half the group unable to see the map.

Prepare the Truth, Not a Script

Write the scenario's core situation as a short causal chain. Start with the antagonist's objective, then describe what they have already done, what they will do next, and what will happen if no one stops them. This is the spine of the scenario.

For example:

A collector has bought a forged family diary to locate a real object hidden in an abandoned estate. They have hired local criminals to keep people away. If they find the object before the next storm, they will attempt a rite that draws an inhuman presence to the house.

That is enough to generate locations, witnesses, evidence, and escalation. You do not need to decree that the investigators must visit the library, then interview the caretaker, then enter the estate through the basement. They may inspect the sale, follow a criminal, contact a relative, break into the collector's office, or set a watch on the property. The truth remains stable while the route changes. The Scenario Design Guide turns this paragraph into a full scenario structure.

For each likely location, prepare three short notes:

• What is visibly happening here?
• What clue, person, or consequence can the investigators find?
• What changes if they spend time here or make noise?

This leaves room for improvisation without abandoning coherence. If the group skips the library, move a key piece of public information to a newspaper archive, a former employee, a police file, or a conversation with someone who has reason to know it. You are making the same truth discoverable through the world, and the scenario still means what it meant.

Build an Investigation That Moves

An investigation needs redundancy. A single clue behind a single successful roll is fragile design, especially in a game where failure and pushed rolls are supposed to matter. The Keeper Rulebook says the same thing plainly: avoid a single concealed clue that unlocks the whole adventure, and make important information reachable in several ways. For every conclusion the group must reach, prepare several paths toward it.

Use a clue map built from questions rather than scenes:

The paths should use different approaches. Let Library Use find a record, social skills gain access to a witness, Spot Hidden reveal a physical trace, and a contact supply an interpretation. This prevents one investigator from monopolizing the case and lets a failed approach redirect the group rather than stop it.

Failure should complicate the investigation without erasing its core evidence. A failed Library Use roll might cost time, attract a librarian's suspicion, produce an incomplete answer, or require a riskier source. A failed social roll might make a witness demand proof, call a rival, or speak only after the investigators solve a smaller problem. The group still has a direction; the cost makes that direction more urgent.

Present Horror Without Withholding Agency

Horror comes from implication, uncertainty, vulnerability, and the cost of understanding. The Horror Atmosphere Guide covers how to build it: contrast before dread, one wrong detail rather than five, selective sensory work, and Sanity used as consequence rather than decoration.

The part that belongs here is the second half of the heading. Make the unknown legible enough to act on. The investigators may not know what the entity is, but they can know that it avoids light, follows a sound, appears after a certain signal, or is focused on a particular victim. Those observations turn terror into tense decision-making.

A group given atmosphere but no usable detail has nothing to decide, and a scene with nothing to decide is a cutscene however well it is described.

End Scenarios With Consequences

Close the Session Cleanly covers where to stop for the night. This is the other kind of ending: what happens once the scenario's central question has finally been answered.

A satisfying ending answers the scenario's central question, even if the answer is bleak. The group may stop the rite, rescue the victim, expose the fraud, contain an object, survive with proof, or realize too late that they solved only one part of a larger problem. They should be able to see what their decisions changed.

After the climax, take a short denouement. Establish what happened to key NPCs, what evidence remains, what authorities or enemies believe, and what the investigators paid. Ask each player for one brief image of their character after the case: a sleepless night, a letter they cannot send, a recovered friendship, a resigned decision to continue. This gives the consequences emotional weight without a long epilogue.

Do not nullify success just to maintain a grim tone. A tragic game can let investigators save someone. A campaign can let them win a case while revealing a larger danger. Defeat means something when it follows from choices and pressure, and very little when every conclusion is overwritten by a last-minute twist.

Record the unresolved threads: a missing person, a name in a ledger, a damaged relationship, a destroyed location, or a contact who now owes the group a favor. These threads are the safest way to grow a campaign because they come from play rather than from a prewritten demand that the same characters return.

How to Run Call of Cthulhu 7e

Preparation gets you to the table. This section covers what to do from the moment players sit down until the session ends: premise, scene framing, rolls, clues, conversations, split parties, action, improvisation, and the close. Use it for a first one-shot, the opening session of a campaign, or a game that has stalled into rules debate and repeated clue searching.

The most useful Keeper skill here is knowing which answer gives the players a meaningful next decision. If investigators search a room and find a diary, the diary should point somewhere, change their view of someone, establish a deadline, or make a current plan more dangerous. If it only proves that your notes contain more notes, the group will keep searching until the session loses momentum. When in doubt, give a clear fact and ask, "What do you do with it?"

Before Anyone Rolls

Your preparation should fit on a working page. You do not need a novel, a scene-by-scene script, or a contingency plan for every possible player action.

Give Players a Usable Premise

Players build better investigators when they know what kind of game is beginning. Give them a premise before investigator creation: enough to choose relevant skills, relationships, and reasons to act without spoiling the hidden threat.

A usable premise contains four facts:

1. When and where are we? Give the era, region, and immediate social environment.
2. What has gone wrong? A disappearance, request, death, discovery, threat, or obligation creates a first question.
3. Why are these investigators involved? Give a shared employer, relationship, profession, place, event, or mutual concern.
4. What kinds of action fit? Mention whether the scenario expects city investigation, travel, social access, wilderness work, research, or immediate danger.

For example: "It is 1927 in New York. A museum assistant you all know has vanished after cataloguing an estate collection. The museum wants the matter handled quietly before an auction. Expect interviews, records, wealthy patrons, and at least one trip outside the city."

That premise gives a journalist, doctor, antiquarian, driver, lawyer, and socialite different reasons to participate. It also warns players that a purely combat-focused investigator may not get much spotlight until the later danger arrives.

Invite players to connect themselves to the premise, and phrase hooks as questions. Who knew the missing assistant? Who has worked with the museum? Who has a reason to distrust the auction house? Replace an unimportant prepared name with a player-provided connection where it fits. A hook the player chose usually holds up better than a relationship assigned after the sheet is complete. The Keeper Rulebook suggests the same move: players often invent relatives, colleagues, and enemies during creation, so note those backstory details for use in the scenario.

Use pregenerated investigators when time is short, such as at a convention. Keep each backstory short, put required knowledge in visible bullets, and make sure every investigator has a reason to care and a chance to contribute. Do not hide the one essential fact in a paragraph someone reads only after the scene has moved on.

Open the First Scene With Pressure

Do not open with, "You are in a town. What do you do?" Open with a change that requires attention. A caller asks for help. A body is found. A deadline is announced. A witness arrives shaken. A routine errand produces an impossible detail.

Make the first scene answer three questions quickly:

• What is wrong right now?
• Why do these investigators have reason to respond?
• What can they do first?

The first action needs to be legible, and it can be perfectly safe while it is. A missing-person report gives names, a location, and a timeline. A strange object gives a physical clue, an owner, and a reason to research. A distraught contact gives the group a witness to question and a deadline to take seriously.

Start in the middle of a relevant situation when it helps. Investigators can already be at the crime scene, on the train to the town, waiting for the client, or answering a disturbing phone call. This avoids spending the first hour arranging a meeting the premise already established.

Frame Every Scene Around a Question

Every scene should have a question that can change the investigation. If there is no question, summarize the travel, routine research, or social transition and move to the next meaningful choice.

Useful scene questions include:

• Will the witness tell the group what they saw before someone interrupts?
• Can the investigators determine who sent the letter?
• Will the group get into the archive without alerting the curator?
• Does the strange residue connect the two locations?
• Can the group leave with the evidence before the building is searched?

State the immediate situation, describe obvious details, and ask for intent. "The clerk is defensive, the appointment book is open behind the counter, and you hear someone approaching through the back room. What are you trying to accomplish?" This gives players agency without forcing them to guess which skill you expected.

End the scene when its question is answered or its pressure changes. Once the witness has given usable testimony, summarize the rest of the conversation instead of requiring fifteen more minutes of roleplay. Once a search has yielded its key evidence, move to the decision about what to do with it. If the table is productively debating, let it breathe. If the group keeps repeating the same plans, introduce a consequence that follows from existing pressure: the shop closes, a contact calls back, an enemy acts, or a deadline gets closer. The Keeper Rulebook gives similar advice for stalled discussion: bring in a desperate NPC asking for help, talk through the odds of each plan, or offer an Idea roll.

Describe First, Roll Second

Players make choices through the information you give them. Describe the ordinary facts before asking for a die roll: exits, people, obvious danger, objects, sounds, weather, time pressure, and what a character would reasonably recognize from their background. Clues are not rewards for reading the Keeper's mind.

Then ask what the player wants. A player who says, "I search the office," may want the missing contract, a hiding place, evidence of a visitor, or an escape route. Those are different actions. Once you know the goal, decide whether the outcome is certain, impossible, uncertain but low-stakes, or uncertain with meaningful consequences.

This is a fast diagnostic, not permission to turn every scene into a negotiation over dice. Rolls are valuable when they create uncertainty with consequences. They are noise when the group needs a basic clue, has unlimited time and no risk, or cannot meaningfully respond to failure.

The Keeper Rulebook splits clues into two kinds, and you decide which one applies before any dice are rolled. Obvious clues are things the players need to know; they find them without rolling. Obscured clues help but are not essential, so either result keeps the game interesting, and a missed roll creates tension because the players know they missed something. The book's example: dust marks show a book was moved (obvious), while a single hair in the dust may go unnoticed (obscured). It also notes that a thorough search can earn an automatic Spot Hidden success.

So give the basic fact freely when the evidence is plain: the door has been forced, the room smells of seawater, the ledger is missing a week, or the witness is terrified of a particular name. A roll can reveal who forced the door, why the water matters, which transaction was removed, or what makes the witness afraid. If an investigator searches the obvious desk where the scenario's central letter is hidden, they find the letter. A successful Spot Hidden can reveal the hidden compartment, the second letter, the fingerprints, or signs that someone searched it first. The game stays investigative because players decide where to look and what the evidence means.

Use handouts with purpose. A document, photograph, map, receipt, or clipped article is most useful when it creates a question the group can discuss. Give it an inconsistency the players can act on: a date that cannot fit, a name that appears twice, a missing page, an address, a mark, or a phrase a witness repeats. A handout that only repeats your narration is decoration. For a remote game, a clear image or a short transcription is usually better than a dense page the group cannot read together.

Choose the Right 7e Procedure

Call of Cthulhu 7e has procedures that make risk visible. Use them when they clarify the scene, and skip them when a player has merely mentioned a skill.

Standard skill rolls. Use a skill roll when an investigator attempts something uncertain and consequential. State the relevant skill and difficulty before the roll when the character could judge the task. A locksmith can tell that a lock is unusually difficult; a researcher can recognize that a document is fragmented; a climber can see the rain making a wall dangerous. Success levels let you distinguish a competent result from an exceptional one, so use them to add speed, quality, safety, detail, or advantage. Tie Hard and Extreme difficulty to the fiction, never to a wish that the group spend longer on the scene.

Pushed rolls. Pushing gives the player a second and final attempt, and the rulebook puts the burden on the player: when someone asks to push, ask what their investigator does to push the situation. More time, more force, a riskier method, a questionable contact, a compromised safeguard, or a willingness to draw attention all qualify. Foreshadow the likely consequence before the dice decide it, because a failed pushed roll lets the Keeper bring serious trouble. Combat, chase, Sanity, and Luck rolls cannot be pushed, and neither can opposed rolls. A failed Library Use roll might become a pushed attempt when the investigator stays after hours, reads restricted material, or calls a person they hoped never to involve. Failure then follows from that escalation: they misfile a record, alert a rival, damage a source, or learn the answer in a way that makes them a target. The Pushing Rolls Guide explains the player side.

Idea rolls and stalled play. The Idea roll is the book's tool for getting a stuck investigation back on track. Players usually propose it, the player whose investigator has the highest INT rolls, and the Keeper picks the most important clue they missed. The group gets that clue either way; the roll decides how. A win delivers it cleanly, woven into the story, and the Keeper may move time forward but should not increase the danger. A loss gives the Keeper permission to drop any of the investigators into the thick of things, starting the scene in the worst situation in which they could realize the clue. Difficulty runs backward from what you might expect. It is Regular if you never mentioned the clue, since the players could not act on something they never heard. It is Hard if the clue was mentioned but not stressed, and that is the default. It is Extreme if you made it clear or the players discussed it themselves. Before offering the roll, restate established facts in plain language; sometimes that alone gets the table moving. If a group never asks for Idea rolls but has lost direction, the book suggests closing the chapter, running the development phase, and opening the next part with an Idea roll.

Sanity checks. Call for a Sanity roll when an event is shocking, horrible, or reality-altering. Describe the experience clearly enough that the player knows why it matters, then resolve the result and its immediate effect. Do not turn every gloomy room or suspicious stranger into a Sanity check; frequent low-impact calls make the procedure feel like ambient damage. The book also caps repeated losses to the same kind of monster: once an investigator has lost as much as that monster's maximum Sanity loss, further sightings stop costing points for a reasonable interval. See the Sanity System Guide and Sanity Checks Guide for the detailed procedures.

Keep Clues Moving

The investigation should survive missed rolls, wrong guesses, and unexpected routes. During play, ask which existing clue can logically appear through the group's chosen method.

Sort information into three categories:

Core clues should not disappear because of one failed roll. Supporting clues reward competence. Advantage clues make the group feel clever without making the scenario depend on perfection.

When the players are out of ideas, separate a character problem from a player-information problem. If the investigators have enough evidence but the players have forgotten it, summarize the known facts or offer an Idea roll. If the players never received a necessary connection, the Idea roll can still deliver it at Regular difficulty, or you can place another clue on the route they are already following. Do not protect a mystery by requiring the table to remain confused.

When the group follows a false theory, let the world answer honestly. A red herring may be a real complication, a partial truth, an antagonist's deliberate misdirection, or a lead that proves irrelevant. It should eventually yield a fact that redirects play. Do not make the group spend an entire session being wrong while you wait for them to return to the exact location you prepared.

Run Conversations as Investigation

NPC conversations are scenes with information, pressure points, and risk. Before an interview, decide what the NPC knows, what they want, what frightens them, and what they will do if treated badly. Then let the players choose an approach.

Not every conversation needs a social roll. A cooperative witness can state what they know. Use a roll when the group is trying to change behavior, overcome a real obstacle, identify deception, or gain something beyond ordinary cooperation. Name the cost of failure: the witness becomes defensive, wants proof, calls someone else, ends the meeting, or shares information later with a rival.

Keep the social skills distinct. The Keeper Rulebook defines them this way:

• Charm uses warmth, flattery, or attraction to compel someone, but not in a way completely contrary to their normal behavior.
• Fast Talk is quick verbal trickery. Its effect is always temporary, and the target usually realizes they were conned after the fast talker leaves.
• Intimidate uses threats, force, or psychological manipulation to frighten or compel, and may leave you with an enemy.
• Persuade uses reasoned argument and takes at least half an hour. For speed, the book points you to Fast Talk instead.
• Psychology lets an investigator form an idea of another person's motives and character. The Keeper may roll it in secret. It gives an impression and does not force anyone to tell the truth.

Charm and Fast Talk can each be opposed by the same skill or by Psychology, and Intimidate by Intimidate or Psychology.

If the group makes a good argument or presents decisive evidence, let that change the situation even without a roll. The dice resolve uncertainty; they should not erase effective roleplay. Likewise, a clever speech does not automatically convince someone whose fear, loyalty, or interest makes agreement implausible. Explain the obstacle and let the players find another angle.

Manage Split Groups and Spotlight Time

Investigators split up because horror fiction teaches them to make regrettable choices. You do not need to forbid it. You do need to keep any player from waiting so long that they stop participating.

When the party separates, establish exactly who is where. Run one short scene until it reaches a decision, danger, discovery, or cliffhanger, then cut to the other group. Tell inactive players when you expect to return to them. A map, tokens, or a quick written list prevents accidental omniscience when investigators are in different places.

Ask questions directly to quieter players. "What does Dr. Bell notice while the others argue?" "Who is watching the street?" "How does your investigator react to this evidence?" You are not forcing a performance, only reminding everyone that their investigator is in the scene.

Do not let one player dictate another investigator's actions. They may suggest a plan, but the player of that character decides what they do. If the group needs a rapid decision, ask each affected player for a concise intention and resolve the actions in a clear order.

Handle Combat, Chases, and Sudden Danger

Action scenes should serve the investigation. A chase can stop a witness from escaping, a fight can buy time to secure evidence, and a break-in can force the group to choose between the object and an injured ally. The action changes the problem instead of replacing it with routine combat.

Before structured action begins, make three facts clear: what each side wants, what exits, obstacles, and routes exist, and what will end the scene. The group may need to escape, protect a witness, retrieve evidence, prevent an alarm, delay a pursuer, or survive long enough for a vehicle to start. Players need enough environmental information to choose between negotiation, a maneuver, a retreat, a chase, or direct force.

Use the combat rules when turn order, range, attacks, maneuvers, injuries, and competing actions matter. Use the chase rules when both sides are trying to gain or close distance along a route where obstacles and choices matter. A short, obvious exchange can be resolved quickly so table time goes to the choices around it. The Combat and Chases Guide and When to Fight, When to Flee Guide cover the player-facing survival framework, and the Major Wounds Guide explains why violence has consequences beyond the current scene.

Keep the objective visible. A player who says, "I hold the door while she gets the file," has given you a clear scene to run. A group still trading attacks after the file is safe and the exit is open may need a reminder that leaving is an available victory.

When an inhuman threat resists ordinary violence, communicate that quickly through the fiction. Let the group see that bullets are ineffective, that a barrier changes the creature's behavior, or that a human cultist wants something different from the entity it serves. A monster does not need to be a fair fight, but it should offer observations that lead to choices.

Do not use action as a punishment for a successful clue. If an enemy reacts to the investigators' progress, telegraph the reaction and let the group respond. Surveillance, a warning, a threat to a contact, or someone removing evidence all work better when the investigators can see the danger arriving and decide what they are willing to risk.

Improvise Without Losing the Truth

Players will ask about locations you did not prepare, pursue a minor name, ignore an obvious door, or turn a casual question into the session's central theory. That is normal. The Keeper Rulebook recommends answering with "Yes, and..." or "Yes, but..." instead of a flat no. Do not punish curiosity by making unplanned actions empty. Improvise from the scenario truth.

Ask four quick questions:

1. What would be true here if the scenario's underlying events are true?
2. Which NPC, clue, or consequence could plausibly connect this action to the investigation?
3. What does the group risk by pursuing this approach now?
4. What changes in the world after they do it?

You can move the delivery of evidence without changing its meaning. If the group does not visit the archive, a retired clerk may hold the same public record. If they watch the suspect instead of interviewing them, they may see the meeting the interview would have revealed. If they approach the antagonist too early, the antagonist may offer a lie, a threat, or a deal that reveals their priorities.

Do not invalidate a clever solution because it skips a prepared scene. If the plan follows from established facts and the group accepts the risk, let it work. Add reasonable consequences if it creates them, but do not quietly restore the missing obstacle because you wanted to use it.

Close the Session Cleanly

End a session at a natural stopping point: a revelation, a consequence, a decision, an escape, the end of a major scene, or a new threat the group can anticipate. Do not run past the group's energy to reach the next page of notes.

For a one-shot, build toward a beginning, middle, and end with escalating pressure, and reserve time for the central confrontation and a brief aftermath. Remind the group of the current objective if necessary, summarize established evidence, and cut lower-priority detours when time runs short. A smaller but complete ending beats a sprawling investigation that stops before the group can act. The rulebook's option when time is nearly up: fast-forward, tell the players what they learned and where it leads, and open the final scene in the middle of the action, possibly using an Idea roll to decide whether they start in a good position.

For an ongoing campaign, end with one clear item to carry forward: a discovered name, a changed relationship, an approaching deadline, a dangerous object, an enemy response, or a question the players want answered. A cliffhanger works when it grows naturally from the session, but it does not replace giving the group enough information to act next time. Take a few notes while everyone is present:

• What did the investigators learn?
• What did they change or damage?
• Which NPCs now know about them?
• What did they promise, take, lose, or leave behind?
• What is the most immediate next pressure?

Ask for a little feedback, especially in a new group. What felt tense? What felt unclear? Was the pace right? Did a theme land badly? A few answers calibrate the next session better than guessing from your own notes.

NPC Creation: People Who Produce Choices

An NPC is a person who can make a choice in front of the investigators. They want something, avoid something, know something, and will respond when the investigators disturb their day. Build those moving parts first. Wardrobe, accent, and tragic childhood can follow when they make that choice clearer.

This keeps preparation sane. Call of Cthulhu investigators tend to ask the wrong person the right question, visit a location early, threaten a witness you expected them to charm, or decide that the harmless landlady is the real villain. A useful NPC brief gives you enough to react without forcing the group back onto a prewritten track. Treat witnesses and allies as seriously as antagonists; a scenario flattens when every non-player character is either a clue dispenser or a villain waiting to be unmasked.

The best Call of Cthulhu NPC makes the investigators choose between two reasonable priorities. An archivist who hides a ritual ledger because exposing it would destroy her late father's reputation can be persuaded, pressured, protected, deceived, or proven wrong. An archivist who is simply secretly evil usually produces one scene, one reveal, and a fight. Cosmic horror already supplies plenty of monsters. Let your human cast supply the difficult decisions.

NPCs Are Choices, Not Lore Dumps

Before writing a name, decide what this person does to the scenario. A good answer is an active verb: delays, warns, hires, conceals, bargains, misdirects, follows, protects, escalates, or breaks. "Knows about the cult" is not a function until you know why they will or will not share that knowledge.

Not every NPC needs to be elaborate. Every important interaction does need to be elastic. When investigators ask an unexpected question, you should be able to answer from the NPC's priorities instead of consulting a script. In investigation play, an NPC's attempt to conceal information can then become a new lead rather than a dead end.

Use this test: if the investigators never met this NPC, what would the NPC do next? If the answer is "nothing," the character may only be scenery. That is fine for a background extra. For a witness, rival, patron, or suspect, give them an agenda that proceeds with or without the party.

NPCs should support the investigators' story rather than replace it. A retired detective may know how to read a crime scene but should not solve it for the players. A powerful patron may open a door but should create a debt or complication in return. The Keeper Rulebook adds a useful reminder: NPCs do not know everything the Keeper knows. Let them make mistakes, misunderstand, and fail. Cultists are only human.

The Five-Part NPC Brief

For any scene-level NPC, write five short answers. One sentence each is enough.

Take Dr. Evelyn Marsh, a hospital pathologist who saw an impossible wound on a corpse. Her goal is to protect the hospital from scandal. Her fear is that the family will accuse her of incompetence. Her resource is access to the body and the original report. Her limit is that she will not alter official records. Her change trigger is proof that someone else has already taken an interest in the case. You can now run her through charm, intimidation, persuasion, bureaucracy, burglary, or simple empathy without deciding in advance which one is correct.

For quick prep, the first four elements are enough: a newspaper editor who wants to keep a scandal out of print (goal), whose sibling attended the cult meeting (fear), who controls the morgue files (resource), and who will not betray a friend without proof they are in danger (limit).

The five parts are play handles, not a personality test. Add a visual detail or mannerism only when it helps the table remember a handle: Marsh keeps checking the corridor because she fears being overheard, and she uses precise medical language because professionalism is her shield.

For a key NPC, add one contradiction. The honest police captain accepts favors from a local industrialist because his precinct needs equipment. The smuggler refuses to traffic in people because a sibling disappeared years ago. Contradictions make a person feel human and create choices that cannot be reduced to a single moral label.

Start With Function, Then Person

Scenario prep is easier when you begin with the role an NPC plays in the investigation. These functions are reminders to give a scene a job, and a person can hold more than one.

A gatekeeper can also be a witness. A suspect can become an ally after the investigators reveal the actual danger. That reuse makes the cast feel connected and keeps the investigation from becoming a row of isolated doors.

NPCs should change the group's options. The police detective can close a location or give the investigators a short window. The frightened caretaker can provide a key but insist on coming along. The rival scholar can translate a page and then demand access to the object. These complications work because they create decisions with consequences.

Do not let a role prove moral character. The official who refuses access may be conscientious, frightened, compromised, or simply busy. The occult bookstore owner may be an exploiter, a careful scholar, or the one adult in town who understands the danger. The Keeper Rulebook warns directly against portraying authorities as uniformly corrupt fools; the game assumes human civilization is worth saving, so some authorities should be shown to be worth the struggle. Avoid making every NPC deceptive, too. If every witness lies and every contact is secretly a cultist, players learn that social play is a trap. Reliable people give betrayal its weight when it belongs in the story.

Relationships, Leverage, and Secrets

An NPC becomes memorable when their choices affect someone else in the cast. Draw a simple relationship map for key characters, with arrows labeled owes, fears, protects, blackmails, envies, trusts, or needs. Three strong connections are usually more useful than ten isolated biographies.

Relationships create pressure points, but use them fairly. The investigators should be able to notice the connection, learn about it, and decide whether acting on it is worth the cost. "The butler suddenly confesses because the plot needs it" is not leverage. "The butler admits he burned a letter once the investigators promise not to involve his employer's daughter" is a decision with stakes.

Secrets need a route into play. For each significant secret, prepare at least two ways the group might discover it. One can be social: an argument, a lie caught in conversation, an old friend who knows too much. The other can be physical or procedural: a bank record, a damaged photograph, a hotel register, a missing page, a witness at a second location.

Separate a secret from a clue. "Mabel caused the fire" is a secret. "Mabel's glove is trapped behind the fuse box" and "the insurance clerk remembers her unusual claim" are clues. If you only write the secret, the scenario's next step stays inside your own head.

An NPC's resource should also have a cost. A librarian can obtain restricted material, but only after hours and only if the investigators keep their voices down. A union organizer can introduce dockworkers, but expects the group not to bring police attention on them. A wealthy patron can pay for travel, but will ask for a report and may object to reckless destruction of evidence. Costs make help feel like a relationship instead of a button the party presses.

Roleplaying Hooks That Move the Scenario

A roleplaying hook is a pressure point that invites play. "Speaks with a stutter" may help you portray someone, but it does not tell you what they want from the investigators. "Needs the investigators to locate a missing brother before the police notice his connection to a labor dispute" gives the table something to negotiate.

Give every important NPC two or three hooks. A useful hook has an ask, a risk, or a trade.

Hooks also tell you how an NPC enters a scene. The constable asks why strangers are examining a crime scene. The scholar sends a worried telegram. The smuggler's runner appears at the hotel with an offer that expires at midnight. These moves create momentum while leaving the investigators free to respond however they like.

When an NPC wants something in return, be plain about the exchange. You can conceal the NPC's deeper motive, but the immediate bargain should be legible: access to the morgue in exchange for discretion, a name in exchange for a favor, an alibi in exchange for silence.

Use Backstory Hooks With Permission

Investigator backstories are a source of motivation. Do not treat them as ammunition to confiscate from players. During preparation, identify one or two hooks you may use: a meaningful relationship, debt, professional reputation, old fear, unfinished case, inherited object, or place tied to the character's past. Ask the player what is in bounds before turning it into danger.

A good hook makes the investigator choose. Their former mentor asks for help but may be compromised. A relative has purchased the wrong book. A professional contact wants the case buried. A character's old research proves relevant, but publishing it would damage someone they care about. The hook brings the investigator closer to the scenario without deciding their response.

Do not repeatedly punish the same player for writing a detailed backstory. Spread attention around the group and let hooks produce help as well as threats. A sibling can offer a place to hide. A former colleague can identify a substance. A trusted friend can become someone worth protecting without automatically becoming a hostage.

Prepare NPCs in Tiers

Match preparation to the chance that the NPC will recur or oppose the investigators.

• Background NPCs get a role and one telling detail. "Night clerk, exhausted, wants his shift to end." "Dockworker, protective of her crew, keeps a brass whistle on a cord." They color the setting and can become more important if the players attach themselves to them. The Keeper Rulebook recommends keeping a list of names ready, since fumbling for one tells players the character does not matter.
• Scene NPCs get the five-part brief plus one hook. They control a lead, complication, or decision. The archive clerk who protects a particular file needs a reason to resist, a reason to bend, and a concrete thing at stake, not a childhood timeline.
• Key NPCs, meaning patrons, recurring allies, principal suspects, and antagonists, add relationships, two discovery routes for each important secret, and an escalation plan. Write what they do if ignored, what they do when confronted, and what pushes them toward their worst choice. Give them a private life only where it can affect play.

A background NPC can graduate when the players become interested. Add a goal and a resource after the fact, consistent with what the table already knows. Developing a person in response to player attention is collaboration with the emerging scenario, not cheating.

NPC Stats Only When Needed

Most NPC scenes do not need a full character sheet. If a person is not likely to resist an investigator with a roll, their statistics are irrelevant. Play the conversation. Decide what their goal makes them reveal, deny, request, or do. Roll only when there is meaningful uncertainty and a consequence for failure.

When an NPC can oppose the investigators, record only the characteristics and skills that could matter in the prepared scenes. A guard may need Fighting, Firearms, Spot Hidden, and perhaps a social skill. A librarian may need Library Use, Occult, and Credit Rating only if those values could change play. Filling a complete sheet for every witness creates paperwork and tempts you to solve scenes with numbers instead of motives.

For improvised NPCs, the Keeper Rulebook gives a quick method. When an investigator acts against an NPC, look at the NPC skill or characteristic that would oppose it. If that value is 50% or higher, set the investigator's roll at Hard; if it is below 50%, set it at Regular. Reserve Extreme for exceptional circumstances, since few NPCs have even one skill of 90% or more. You can then adjust for motivation: the book's example raises the difficulty to Extreme for a cultist who is determined not to talk, even though his Intimidate and Psychology are both below 50%. If a fight breaks out and you have no numbers, assign combat values on the spot: about 25% for an unskilled average person, 40% for a barroom brawler or hired thug, and 70% for a trained killer. A high skill is a promise about the fiction, so the investigators should be able to see why this person is unusually good at something.

First impressions. You decide how an NPC reacts based on the fiction. If you want a random element, the rulebook offers a concealed 1D100 roll against the investigator's APP or Credit Rating, whichever the NPC cares about more; a result below the value means a positive reaction. Its example uses Credit Rating for a desk sergeant who is more impressed by wealth than looks. Do not turn every introduction into a reaction roll; a known history, an obvious threat, or a well-made argument can matter more than either number.

Combat is a behavior choice. If an NPC becomes involved in combat, decide what they are trying to preserve. Someone who wants to escape will often dodge, flee, call for help, or create an obstacle. Someone cornered and determined to stop the investigators may fight back. Choose actions that show the NPC's goal, fear, and limit, even when a stat block suggests something more efficient. For the procedures, use the combat rules guide. For social pressure that risks escalating, the pushing a roll guide is a reminder that consequences should advance the fiction.

Run NPCs Under Pressure

Prepare an escalation ladder for key NPCs. It keeps a confrontation from stalling at repeated refusal and gives you a fair response when the investigators apply pressure.

Treat this as a menu of honest reactions. Let the investigators' approach change the order and outcome. If they offer safety before making threats, a frightened witness may cooperate at once. If they publicly humiliate a proud official, that official may become an obstacle even after accepting the evidence.

Keep the person distinct from the information they hold. A frightened witness should do something more interesting than become unavailable. They might insist on meeting elsewhere, demand a promise, give an incomplete warning, contact the investigators later, or run to someone who creates the next scene. A failed social interaction can produce trouble without erasing the lead.

Use short physical beats to reveal pressure: a hand moves toward a locked drawer, a sentence is corrected too quickly, a coat stays on because the person plans to leave, a cup goes untouched. These details work better than announcing "this NPC is nervous," and they give observant players something to pursue.

Common NPC Mistakes

The clue vending machine. An NPC who exists only to answer one question disappears from memory. Give them a reason to care about the answer. Even a helpful witness may want reassurance, protection, recognition, or time.

The Keeper's favorite. Do not protect a beloved NPC from consequences the investigators have earned. They may mistrust, embarrass, arrest, recruit, abandon, or accidentally kill someone you enjoy portraying. Let that matter.

The single-point secret. If the entire scenario requires the group to notice one lie from one person, the NPC is a bottleneck. Put evidence in relationships, locations, documents, and consequences as well as dialogue.

The universal corrupt authority. An obstructive official is useful when the obstruction has a reason. A principled archivist, overwhelmed patrol officer, or cautious doctor can create friction without turning the world into a cartoon.

The full sheet before the first sentence. Numbers are not personality. Start with what the NPC wants, then add only the stats that could resolve a real conflict.

The one-note performance. An accent, stammer, or catchphrase does not replace an agenda. Avoid turning identity, disability, ethnicity, or class into a gimmick. Give every NPC dignity, specific interests, and the ability to surprise the table.

NPC Worksheet

Copy this for any NPC likely to get more than a few minutes of table time.

Name and role:
First impression:

Function in this scenario:
Goal right now:
Fear:
Resource or access they control:
Limit they will not cross:
Change trigger:

What they want from the investigators:
What the investigators can reasonably want from them:
Relationship connections:

Secret:
Discovery route one:
Discovery route two:

Roleplaying hook one:
Roleplaying hook two:
If ignored:
If confronted with evidence:
If frightened or cornered:

Only needed mechanics:

Use the first impression as a practical cue. "Clean cuffs despite the rain" is easier to play than three paragraphs describing a face. The final check is simple: can this person change the investigators' plans, and can the investigators change this person's plans? When both answers are yes, you have an NPC rather than a delivery system.

Campaign Pacing: Discovery, Danger, and Rest

Pacing is the Keeper's control of when scenes begin, linger, cut away, and end, and at campaign scale it is the rate at which the game changes the investigators' understanding, options, safety, and obligations. A good session alternates among discovery, planning, danger, and recovery. Constant danger becomes noise; constant research becomes inventory management.

A campaign needs a rhythm: a problem arrives, the investigators find their footing, the situation tightens, consequences land, and the group gets enough room to understand what changed before the next pressure arrives. Without that rhythm, every session feels equally urgent until urgency stops meaning anything. Call of Cthulhu campaigns often begin with ordinary people investigating an alarming local event and end with those same people deciding what they are willing to lose to prevent something worse. Good pacing makes that change visible. Most of what follows is tablecraft rather than rules.

A Session Rhythm That Works

Use this rhythm as a practical default for a single session:

1. Open with movement. Present a troubling fact, request, consequence, or deadline that makes a response necessary.
2. Let the group investigate. Give them time to interview, search, compare evidence, and form a theory.
3. Change the pressure. Introduce an approaching deadline, hostile response, disturbing manifestation, lost opportunity, or cost of delay.
4. Allow a decision. Let the group choose a route, ally, risk, or objective with the information available.
5. Give them an aftermath. Resolve the cost, reveal what changed, and let them plan before escalation resumes.

Slow down for a revelation, a hard moral choice, a frightening description, or a player conversation that matters. Speed up routine travel, repeated searches, and logistics that no longer create choices. If players are actively planning, give them room. If they keep repeating the same theory without new evidence, change the situation with pressure that follows from the scenario:

• The missing person misses another deadline.
• A contact is suddenly unavailable.
• An antagonist learns the investigators are involved.
• The weather, transport, or local authority changes the available routes.
• A new clue contradicts the group's safest assumption.
• A witness needs help now rather than later.

The investigators should be able to look back and see why their delay, visibility, or choices mattered. That fairness makes escalation frightening instead of arbitrary.

What Campaign Pacing Controls

Pacing covers more than how much happens in a session. A focused three-hour conversation with a witness can move a campaign faster than a sprawling chase if it changes what the group believes and what they must do next.

Track four questions after every session:

1. What did the investigators learn?
2. What became harder, more dangerous, or more expensive?
3. What can they now choose that they could not choose before?
4. What needs time at the table before the next escalation?

Keepers skip the fourth question most often. The group may need to check a source, tend an injured ally, argue about an offer, contact a relative, or decide whether to continue. If you interrupt every pause, the campaign becomes a sequence of Keeper commands instead of an investigation driven by player choices.

Pace the information, not just the threats. The players should usually understand the current problem well enough to make a meaningful decision, even when they do not understand the full supernatural truth. Mystery comes from unanswered questions with evidence pointing somewhere.

Build a Campaign From Arcs

An arc is a set of scenarios connected by a question that can eventually be answered: "Who is importing these artifacts?", "What happened to the missing expedition?", or "Why do the same symbols appear in unrelated deaths?" An arc can last three sessions or twelve. It needs a beginning, a middle, a turning point, and a consequence that changes the campaign.

Avoid a single unbroken chain in which every case must be solved in a particular order. Build from linked nodes instead:

Give each scenario its own beginning and ending even when it serves a larger arc. The group should be able to say what they accomplished, failed to accomplish, or changed. A campaign made only of setup makes every session feel like a prologue.

Run two or three active threads at once, not seven. One can be urgent, one personal, and one an optional mystery or resource problem. That gives players real agency about what to pursue while keeping the campaign legible. If every clue points to a different continent, organization, and apocalypse, the group is being asked to sort the Keeper's filing cabinet.

The Keeper Rulebook flags a practical problem with long campaigns: a reason that ties investigators to their first scenario can feel labored if reused, so unrelated scenarios with the same group need a good reason for them to keep getting involved. Recurring threads, patrons, and unresolved consequences supply that reason.

The Four Rhythms to Track

Strong long-form games vary four overlapping rhythms. You do not need a chart, but you should know which one you are raising and which you are letting settle.

Information. New evidence, corrected theories, and reliable conclusions create forward motion. Early in an arc, give broad questions and visible leads. In the middle, let the group learn the opposing force's method, goal, or vulnerability. Near the end, reveal the cost of acting and the cost of doing nothing. Answer earlier questions regularly. An answer can expose a larger question and still feel like an answer: "The murderer was the curator" is progress even when the curator worked for someone else.

Threat. Threat does not have to mean combat. It can be surveillance, social isolation, a deadline, legal pressure, a vulnerable person, an institution closing ranks, or an unsafe environment. Alternate overt threat with implied threat. A note proving someone entered the investigators' offices can be more useful than an immediate attack because it gives the group room to respond.

Emotion. Horror needs ordinary life around it. Plan scenes where relationships, professional duties, pride, money, guilt, and curiosity matter before the next uncanny development. They are what make the horror costly. A character with nothing normal left to lose cannot make the final choice hurt.

Mechanics. Vary what the table is doing: social investigation, document analysis, travel, surveillance, a tense entry, a chase, a fight, recovery. Do not call for a roll merely to change the activity; a well-run conversation or handout can carry a whole scene. When you do call for rolls, make success, failure, and pushing change the situation, as covered in the 7e procedure notes.

Plan a Scenario's Place in the Campaign

Before you prepare a scenario, write one sentence for its campaign job. It may introduce a threat, test a relationship, reveal a method, remove a safe assumption, provide a resource, or force the group to choose between priorities. "It is a cool haunted house" is a setting. "It proves the antagonist can reach the investigators' families" is a campaign job.

Then choose a tempo. This sequence is dependable but not mandatory:

The most useful slow session in a horror campaign is the one where the investigators discover what the last disaster cost them and choose what they will do differently. A breather still has events; the primary threat just is not demanding an immediate reaction. The group may research, make contacts, spend money, repair trust, pursue a personal lead, or decide how to use what they learned. A quiet funeral, a difficult conversation with a patron, or a normal dinner interrupted by one wrong detail gives danger contrast, and contrast is where dread lives. If every game ends with a cult raid, a chase, and a devastating revelation, the table eventually treats all three as routine.

Do not save every major revelation for the finale. A mid-arc turning point keeps a campaign from becoming a long march toward a twist the players cannot influence. Let them learn enough to change strategy: identify a false ally, understand the cost of a ritual, discover a vulnerable location, or realize that their patron's stated goal is incomplete.

Use Revelation Clocks, Not Constant Emergencies

A clock is a short list of events that occur if the investigators do not interfere. It shows that other people act; it is not a punishment for taking time to think. Write events the group can notice, prevent, redirect, or exploit.

An artifact broker may first move a shipment, then silence a witness, then host a private sale, then attempt a ritual at the buyer's estate. Each step creates evidence and choices. If the group stops the shipment, the broker might accelerate the sale. If they warn the witness, the broker may target their reputation instead. The clock advances the situation without declaring the story failed.

Use a clock when one or more of these are true:

• The opposition has a plan that makes sense without the investigators.
• Delay would change who is at risk or what resources remain.
• The group needs a reason to choose between several valid leads.
• You can show the next step through evidence, an NPC action, or a public event.

Do not advance a clock every time players pause to discuss a clue. Announce or imply the pressure, then let them decide how to respond. Invisible deadlines train players to rush randomly or stop asking questions.

Make a clock visible only when the players could plausibly know it exists: a newspaper deadline, a scheduled voyage, a public trial, a recurring astronomical event. Keep private opposition clocks in your notes, but show their effects in the fiction. "The train leaves at midnight" is a decision. "Something bad happens because you took ten minutes" is just frustration.

Protect Investigation and Recovery Time

Investigation needs slack. A group cannot make deductions when every scene begins with an attack, and players cannot care about a new lead if they have had no time to discuss the last one. End some sessions after a reliable discovery or a decision point instead of another interruption, and let the group form theories between sessions.

Recovery is part of the pace too. In Call of Cthulhu 7e, injuries and Sanity effects can last well beyond a single encounter. You do not need to erase them to prevent a death spiral. Schedule periods where the investigators can seek medical care, research safely, tend to practical matters, and decide how much danger they are ready to accept. Recovery scenes make consequences legible and let the group show character.

Use these scenes to surface change:

• An ally asks whether the investigators will take another case.
• A newspaper prints a distorted account of their last success.
• A family member notices that someone has changed.
• A bill, official inquiry, or professional obligation complicates the next step.
• A recovered clue suggests the danger was not fully contained.

The point is to connect survival to a living world, not to punish the group for surviving. A recovery window can introduce the next thread gently, but it should not secretly become an ambush every time. For the procedures behind those scenes, see the Sanity rules guide and the major wounds guide.

Vary the Shape of Danger

Campaigns grow monotonous when every case follows the same order: interview, library, basement, combat, reveal. Repeat a familiar structure only when you want the players to notice a pattern.

Action scenes work best as punctuation. A chase can turn careful investigation into urgent movement when the group knows why catching or escaping matters. A fight should alter the investigation afterward: a witness is hurt, evidence burns, authorities arrive, an enemy recognizes the investigators, or a safe location stops being safe. The chase rules guide covers the procedure; decide first whether the campaign needs the change a chase will cause.

Likewise, do not make every supernatural manifestation louder than the last. Scale can increase through intimacy: the first sign is a strange smell, the next is a friend behaving incorrectly, and the next is evidence that the friend chose to cooperate. Closer is often worse than bigger.

Escalate Consequences, Not Just Numbers

More enemies, larger creatures, or harder rolls are the quickest way to make a campaign feel like it is escalating, and the least distinctive. Escalate what a failure, success, or delay changes.

Early consequences may be private and reversible: embarrassment, a lost contact, a missed lead that can be found elsewhere. Mid-campaign consequences can change relationships or resources: a patron withdraws support, an ally is compromised, an institution starts watching the group. Late consequences should force a values-based decision: save people or preserve evidence, tell the truth or prevent panic, destroy a dangerous tool or use it once more.

Keep consequences proportional to the warning the players received. A choice can be hard without being a trap. If preserving evidence will endanger a witness, make the danger visible enough that choosing the evidence is a real decision. If the group never had a chance to know the cost, the result is arbitrary rather than tragic.

Use recurring costs carefully. A campaign in which every ally dies teaches players not to form attachments. A campaign in which every official is corrupt teaches them to avoid institutions. Let consequences vary: an ally may set a boundary, ask for help, leave town, become a complication, or grow into a more capable partner. Loss is powerful because it is possible, and it loses that power once it becomes automatic.

Bring Back NPCs and Earlier Choices

Recurring NPCs are the campaign's memory. When a person returns, show how prior events changed them. The librarian who once needed protection may now be braver but slower to trust. The constable the group deceived may have lost credibility. The rival who escaped may arrive with useful information and a justified grudge.

Bring an NPC back because they have a current agenda, not because the campaign needs a familiar face. Before reintroducing someone, answer three questions:

1. What changed for them since the investigators last saw them?
2. What do they want now?
3. Why are the investigators relevant to that goal?

Tie callbacks to decisions, not trivia. A detail players forgot from session two is rarely satisfying on its own. A choice they made in session two that changed a person's life is satisfying because it proves the world remembers. Run returning characters with the same brief described in The Five-Part NPC Brief so they stay people with motives.

Use campaign handouts sparingly. A letter, report, clipping, map, or photograph works best when players can interpret it or compare it against their notes. Do not make a handout carry the only essential clue; it should reinforce, complicate, or make tangible information that also has another route into play.

Know When an Arc Is Over

An arc ends when its central question has been answered and the group has made the decision that answer demands. It does not need every loose thread tied off. Leave some texture in the world, and close the question the group spent their energy pursuing.

Before an arc finale, make sure the investigators understand:

• What the opposition is trying to do.
• What will happen if nobody acts.
• What options are available, even if none are safe.
• What each major option costs.
• Which relationships and prior choices are likely to matter.

The finale should test what the campaign has taught. If the arc centered on secrecy, the ending should ask what truth is worth exposing. If it centered on institutions, it should ask whom the investigators can trust and what they will do without official support. If it centered on a dangerous resource, it should ask whether using that resource makes them complicit.

Afterward, play an aftermath scene before introducing a new emergency. You can seed the next thread, but it should feel like an invitation rather than a stolen ending. The next section covers how to run that ending.

Campaign Pacing Worksheet

Use this before starting a new arc, and revisit it every two or three sessions.

Arc question:
Human stake:
Opposition goal:
What changes if investigators do nothing:

Opening case and its campaign job:
Mid-arc turning point:
Arc decision the finale should demand:
Likely aftermath:

Active thread one:
Active thread two:
Optional thread or personal connection:

Current information beat:
Current threat beat:
Current emotional beat:
Current mechanical activity:

Next visible clock event:
Recovery or reflection opportunity:
Recurring NPC who can return, and why:
Earlier choice that can matter again:

The worksheet is not a commitment to a script. Cross out entries when play changes the campaign. Its job is to keep pressure, information, and breathing room in view while you prepare. The best measure of pace is whether the group can look back after several cases and see a clear path of discoveries, choices, costs, and changed relationships. When they can, the campaign has momentum. When they also have enough quiet to care about that path, it has weight.

Separate Tablecraft From 7e Procedures

Some Keeper advice is universal: ask for consent, frame clear choices, keep clues flowing, share spotlight time, and let consequences follow from fiction. Those principles work in any system.

Other decisions are specifically 7e procedures. Use the rules when the system provides a meaningful structure:

This separation keeps rulings consistent. Do not hide behind "the rules" when a player lacks basic information their investigator could see. Do not replace a clear 7e procedure with vague narration when a roll, consequence, or timing rule is what makes the decision meaningful.

Era matters too. The Keeper Rulebook supports both the 1920s Classic Era and the modern day: it tags some skills and occupations as Modern (Computer Use, Electronics, and the Hacker occupation, for example), lists modern equipment, and includes a modern investigator sheet. Even so, do not assume that a 1920s institution, technology, legal response, or social pressure transfers unchanged to a contemporary game. The Modern Settings Guide covers that adaptation.

Keeper Resources Guide Library

Keeper Session Checklist

Before the session:

• Confirm the table's content boundaries and the intended tone.
• Write the scenario truth, active threat, and what changes if no one intervenes.
• Prepare multiple evidence paths for every conclusion the group must reach.
• Give each important NPC a goal, fear, resource, and limit, plus two routes into every important secret.
• Note only the NPC stats that likely opposition will need.
• Identify one or two permitted investigator hooks and one way each can help as well as complicate.
• Confirm the campaign has a current problem the players can describe, at least one answer to an earlier question, and a visible or discoverable source of pressure.
• Decide which scenes need 7e procedures and which can be resolved through ordinary table conversation.

At the start:

• State the premise, immediate problem, and first available action.
• Remind players of facts their investigators would know from the previous session.
• Put important handouts, maps, or names where the group can reference them.

For each scene:

• Describe obvious facts, danger, exits, people, and pressure before asking for a roll.
• Ask players what they want to accomplish, not only which skill they want to use.
• Call for a roll only when outcome and consequence are uncertain.
• Give core clues through chosen actions; use rolls for depth, speed, risk, or advantage.
• Share attention deliberately, especially when the group splits up.
• End the scene when its question is answered or the situation changes.

When play stalls:

• Restate established facts before adding more mystery.
• Ask what conclusion the group is trying to test.
• Offer an Idea roll to deliver the missed clue.
• Add a consequence or deadline that follows from the active threat.
• Move a clue's delivery to a logical new route if the original route is closed.

After the session:

• Record what the investigators learned, changed, damaged, and left unresolved, including injuries and Sanity effects.
• Update antagonist plans and clocks based on the group's actual actions.
• Note which NPC relationships strengthened, broke, or became useful complications.
• Check that recent sessions varied their activity and left room to research, recover, or argue.
• Prepare the next session's opening around a consequence or unanswered question from play.

A Keeper does not need to be a novelist, actor, rules database, and stage magician at once. Prepare the truth, listen to the players, make the consequences clear, and let the investigation become its own story. When the group reaches the final door, they should know why it matters and still have to decide whether opening it is worth the price.

Where To Go Next

Running a session well still depends on the procedures that fire inside it. The Combat and Chases Guide and the Cthulhu Mythos Guide cover the two moments that most often need a Keeper ruling mid-scene. The Call of Cthulhu 7e Guide Hub collects the remaining guides, player-facing and Keeper-facing alike.