How to Run Your Own Prediction Market Venue: The Complete 2026 Operator's Manual

Launching a prediction-market venue is one project. Running it is a different one. The launch playbook ends at day 30; this manual starts at day 31 and runs through the year-1 to year-2 transition. Most of the operator value created across this category in 2026 will be created by operators who run their venues well after launch — far more than will be created by operators who simply launched first.

This is the operator's manual for that running. It assumes you've already shipped (the launch playbook covers the shipping; the 15-minute walkthrough covers the technical deploy). It assumes you have an audience, real volume, and a small team capable of executing on operator decisions. From there, here's how to run the venue through the first year and into the second.

The 4× figure is the gap between operators who run the venue attentively and operators who run it as a launched product they don't actively manage. The infrastructure is the same; the operator decisions are not.

The three KPIs that actually matter

A running venue generates dozens of metrics, but three of them explain almost everything about the venue's trajectory.

Weekly active traders (WAT). The count of unique traders who placed at least one trade in the past seven days. This is the single most important number for predicting where the venue is in twelve months. WAT growth, sustained, compounds into volume; WAT decline is the leading indicator of operational problems that will manifest in volume metrics within a month.

Median trade size. The median (not mean) notional per trade across the venue. The median is more honest than the mean because a few large institutional trades distort the mean without affecting the audience's typical experience. Median trade size growing over time signals that the audience is gaining conviction and sizing in; flat or declining median size signals that the audience is hedging or losing confidence.

Resolution dispute rate. The fraction of resolved markets that went to dispute rather than the optimistic-oracle clean path (we covered the mechanics in the resolution post). On a healthy venue this number is below 0.3%. Above 0.5%, it indicates contract specification problems that will erode trader trust over weeks. The metric is leading; trust collapse is lagging.

These three KPIs should be visible at all times to the operator. Weekly review of just these three numbers, with the discipline to act on what they reveal, is the foundation of running a venue well. Other metrics matter, but they're mostly diagnostic of these three.

Market creation cadence

The single highest-leverage operational lever is the cadence and quality of new markets. Audiences trade what's available. The pipeline of new markets is what keeps existing traders returning and what attracts new audiences from adjacent segments.

The cadence question is "how many new markets per week?" The answer depends on audience size and vertical, but the patterns that work are narrower than most operators expect.

For a venue with 1,000 weekly active traders in a focused vertical (e.g., macro contracts), the right cadence is roughly 8–15 new markets per week. Below 5 per week, returning traders find nothing new and churn. Above 25 per week, the quality drops because thoughtful contract specification at that volume requires more operator capacity than is sensible.

For a venue with 10,000 weekly active traders across multiple verticals, the cadence is roughly 30–50 per week, distributed across the active verticals. Beyond that, market creation has to be partially automated through templated pipelines.

The discipline is harder than the cadence number suggests. Each new market is a contract specification decision (which data source, which resolution rule, which edge cases), a liquidity decision (is this market mainstream enough to share liquidity, or does it need seeding), and an operational decision (when does it close, what's the support exposure if it goes to dispute). A venue that ships 10 markets per week needs a process for making those decisions consistently.

The pattern that consistently works: a market-creation team of 1–2 people, each with a defined vertical they're responsible for, working from a backlog of audience-suggested and event-calendar-driven candidate markets. Output is reviewed weekly against the dispute rate; if disputes spike on a specific creator's markets, the spec discipline gets tightened.

Listing decisions: the signal-vs-noise problem

A natural temptation as the venue grows is to list more markets, more variety, more verticals. The data consistently shows this is a mistake. The venues that out-perform are the ones that list disciplined.

The framework for listing decisions has three filters every candidate market must pass:

1. Resolution clarity. Can the resolution rule be written in two sentences without ambiguity? If not, the market isn't ready to list. Almost all dispute-rate problems trace back to listings that failed this test at creation.
2. Audience interest. Will at least 15 traders take a position on this market in the first week? This is hard to predict perfectly, but if the operator can't make a defensible argument for why this market matters to the audience, it probably doesn't.
3. Liquidity feasibility. Can the market trade with a spread under 100 bps in the first 24 hours? If the answer is "no, but we can market-make it," the operator is adding cost. If the answer is "no, and shared liquidity doesn't cover this category," the market won't trade well and shouldn't ship.

A market that fails any of the three should not list. Most operators that struggle with retention struggle because they're listing markets that fail filter 1 (resolution clarity) or filter 3 (liquidity), and the audience experiences those failures as venue-level problems even though they're listing-level problems.

Fee evolution over time

The fee schedule we recommended for launch (covered in the fee models post) is a starting position, not a permanent setting. The operator's job through the first year is to evolve the fee based on what the data reveals about audience tolerance.

The three signals that warrant a fee change:

1. Week-12 retention below 35%. The fee is hurting volume more than it's earning revenue. Drop the rate by 20–25 bps and re-test for another 90 days.
2. Median trade size declining over multiple months. Traders are getting cost-sensitive. Drop the rate by 10–15 bps and watch whether median trade size recovers.
3. WAT growing faster than volume per trader. Pricing power exists. The fee can rise 10–15 bps without losing audience.

The cadence of fee changes is also operational discipline. Aggressive operators want to change fees monthly. The data doesn't support monthly changes; the signal in fee-rate behaviour takes 8–10 weeks to stabilise, and faster adjustments inject noise without producing better understanding.

The discipline that consistently works: hold the launch fee for the first 90 days, run one well-considered adjustment at day 90 based on retention data, hold that for 90 more days, and consider tiered fees only after the venue has 6+ months of consistent operation and meaningful institutional flow.

Trader retention strategies that actually work

Retention is the leading indicator of every other metric. Acquired traders who don't return cost almost as much as acquired traders who never trade at all. The retention strategies that consistently produce results are operationally simple, but they have to be done deliberately.

Resolution communication. When a market resolves, the trader should get a notification within minutes (not hours) showing the outcome, their position, and the credited amount. This sounds trivial; many operators get it slightly wrong (delayed notifications, unclear position summaries) and lose 5–8 percentage points of retention as a result.

New-market relevance. The new markets shipped each week should be relevant to the existing audience. A macro-focused audience doesn't get sports markets dropped on them; a sports-focused audience doesn't get political markets. The weekly digest of new markets should feel like the venue "gets" what the trader cares about.

Position-management UX. Traders should be able to see all open positions in one place, close positions in one click, and see their realized and unrealized P&L without navigation. Fragmented position UX is the #1 reason traders churn after their first few weeks; they can't tell how they're doing, so they stop trading.

Educational content for the second tier. The traders who sign up but only place one or two trades in their first month are the ones who need education to convert into real activity. Content (newsletters, tutorial videos, explainer threads) that goes specifically to this cohort roughly doubles their conversion to active traders compared to leaving them alone.

The pattern across all four is that retention is operational quality, not strategic genius. Operators that out-perform on retention are the ones with disciplined operational attention to these specific touchpoints. The strategy is simple; the execution is what differentiates.

Customer support as a retention lever

Support is usually treated as a cost center. On a prediction-market venue, support is one of the most effective retention mechanisms available, and operators who get it right see meaningful retention compounding from it.

The pattern that works:

• First-response within an hour. The first reply doesn't have to resolve the issue; it has to acknowledge it and set expectations. Traders who feel heard within an hour retain at meaningfully higher rates than traders who wait six hours, even if both end up resolved equally.
• Real names from real humans. Support replies signed by named people, not "the support team," produce dramatically better trader trust. The cost of having actual names on replies is tiny; the retention impact is meaningful.
• Issue categorisation that drives product fixes. Every support ticket should be tagged by category (deposit issue, market clarity, position confusion, withdrawal delay, technical bug). Weekly review of the category mix identifies the product gaps that produce support load. Closing those gaps is permanent retention improvement.

Support volume itself is a metric. Healthy: under 5 tickets per 100 weekly active traders. Watch: 5–15. Bad: above 15. A spike in support volume is a leading indicator of either a specific incident (which the on-call rotation handles) or a specific UX gap that the product team should address.

Resolution operations as ongoing work

Resolution is sometimes treated as automated infrastructure that doesn't need operator attention. That's wrong. A running venue with hundreds of markets per month has dozens of edge cases per week — markets where the underlying data is ambiguous, where the source moves, where a major event reframes the resolution question. Each one needs operator judgement.

The operator's resolution operations playbook:

• A defined "resolution-ops" person or rotation. One named person per week is responsible for actively monitoring resolutions, fielding edge cases, and signing off on resolutions that flagged for review.
• A pre-published resolution policy. A document that states how ambiguous cases are handled. Traders should be able to read it before disputes happen, not after. The policy itself doesn't have to be long; it has to be clear.
• Public documentation when an edge case is handled. When the operator makes a non-trivial resolution decision, publishing the reasoning publicly (not just to the affected traders) builds long-term trust. The operators with the strongest resolution reputations are the ones who've been transparent about their reasoning, not the ones who never had edge cases.
• A relationship with the protocol's resolution support. Some edge cases require protocol-level escalation (especially anything involving the optimistic-oracle dispute path). Operators should have a documented escalation path with the protocol provider and use it proactively rather than reactively.

A clean resolution-ops practice is what separates venues with 0.2% dispute rates from venues with 0.6% dispute rates. The protocol-level mechanics are the same; the operator-level discipline is what differs.

Communication during incidents

Every running venue has incidents — payment provider outages, market resolution delays, scheduled maintenance, occasional bugs. The operator's communication during incidents is what determines whether they damage trust or not.

The communication patterns that consistently work:

• Proactive disclosure. Tell the audience about the incident before they ask. A 10-minute Twitter post acknowledging "we're seeing slower resolution on market X, investigating" builds trust. A 24-hour silence followed by "we have an update" destroys it.
• Specific timelines. "We expect resolution in the next hour" is better than "we'll update soon." If the timeline slips, update. The discipline is in being specific even when uncertain.
• Post-incident write-ups. After any non-trivial incident, write a public post-mortem that explains what happened, what was affected, and what's being changed to prevent recurrence. Operators who do this consistently build long-term credibility that survives incidents.
• Personal acknowledgement to affected traders. Direct outreach to traders whose positions were specifically affected. Not auto-emails — actual human messages. This is expensive in operator time and worth the cost.

The pattern across all four is that incidents are opportunities to strengthen trust if handled well, and opportunities to lose it if handled poorly. Operators who treat incidents as PR risks rather than trust events underperform.

Regulatory relationship management

Through the first year of operation, the regulatory relationship in the operator's primary jurisdiction has to mature from "we launched compliantly" to "the regulator knows us, sees our reporting, and is comfortable with our posture." This is harder than most operators anticipate because regulators move slowly and require continuity.

The minimum operator-side regulatory practice:

• Quarterly written submissions. Even when not required, voluntary quarterly reports to the relevant regulator showing volume, trader count, dispute rate, and compliance metrics build a relationship that becomes invaluable when something goes wrong.
• Pre-emptive disclosure of changes. If the operator changes the fee model, the market scope, the audience targeting, or any compliance-relevant parameter, the regulator hears about it before the public does. Surprise is the single thing regulators dislike most.
• A named regulatory liaison on the operator team. Not a part-time legal counsel — a named full-time person who owns the regulator relationship. By the time the relationship matters, the operator wishes they'd hired this person earlier.
• A clean separation between operational decisions and regulator-facing decisions. Operations move fast; regulator-facing changes move slowly. The operator's process should make this explicit so operational improvements don't accidentally violate compliance posture.

Regulatory tightening cycles will happen. The operators who survive them are the ones who built the relationship through the calm period. Operators who skipped relationship management get cut off when the cycle tightens.

Audience growth: the second-year question

The first 6 months of a venue's life are about activation and retention. The second 6 months are about audience growth. The shift is harder than most operators anticipate because the playbook is different.

For audience growth in months 7–12, three channels reliably produce returns:

Referral mechanics inside the existing audience. A trader who's been active for 60+ days is meaningfully more likely to refer another trader than a trader in their first month. Referral programs work better at the second half of the year than at the first. Build the referral mechanic at month 6, not at launch.

Content-driven SEO and audience expansion. By month 6, the venue has enough operational data to write authoritative content about its niche — market analysis, audience behaviour, resolution patterns. That content compounds in search rankings over 12+ months and is a primary acquisition channel by month 18.

Vertical expansion to adjacent audiences. A macro-focused venue at month 8 can responsibly add a sports or political vertical without diluting the macro audience. The expansion brings in audiences from the new vertical; some of them stay on the macro side too.

The pattern that doesn't work at this stage: spending heavily on cold paid acquisition. The unit economics for prediction markets in 2026 don't support paid CAC above ~$30 per active trader, which is roughly 3× cheaper than typical fintech CAC. Operators who try to "grow faster through paid" end up burning capital on traders that don't activate.

Tier introduction: when and how

By month 9–12, most operators have enough volume distribution to introduce tiered fees. The tier structure that consistently works is a three-tier system based on 30-day volume, with tier breaks at meaningful volume thresholds.

Common pattern:

• Standard tier (default). Whatever the launch fee ended up at, e.g., 0.95%.
• Mid tier (>$10k 30-day volume). A small discount, e.g., 0.80%.
• Top tier (>$100k 30-day volume). A meaningful discount, e.g., 0.65%.

The thresholds and the discount sizes should be tuned to the specific venue's distribution. The discipline is to introduce tiers only after the venue has stable operations, to communicate the tier structure clearly, and to grandfather existing traders into appropriate tiers based on their historical activity.

Tiering pre-month-9 is consistently a mistake. The data isn't stable enough; the discounts confuse the audience; the operational complexity outweighs the revenue optimization.

Reporting and analytics

By month 6, the operator should have a reporting setup that surfaces the three core KPIs (WAT, median trade size, dispute rate) plus a structured set of secondary metrics (activation rate by acquisition channel, retention by cohort, support volume by category, resolution time by market type).

The reporting cadence that works:

• Daily. WAT, trades, volume, deposits, withdrawals, support tickets. Operator dashboard.
• Weekly. Cohort retention, market-by-market volume, dispute summary, fee revenue. Operator team review.
• Monthly. Strategic review covering all KPIs, content performance, audience growth, regulatory submissions. Stakeholder review.
• Quarterly. Full strategic review including competitive positioning, vertical expansion decisions, fee evolution decisions, and regulatory relationship status. Board / executive level review.

Operators that don't have this reporting cadence by month 6 spend the second half of the year guessing at metrics they should know precisely. The reporting setup pays for itself many times over.

The year-1 to year-2 transition

By month 11, the operator should be planning the year-2 strategy. The transition from year 1 (launch + activate + retain) to year 2 (expand + monetize + scale) is a real shift and worth being deliberate about.

The year-2 priorities that consistently emerge:

1. Vertical expansion. Add a second or third vertical with the operational discipline that month 6+ proved works. Start narrow inside the new vertical and expand from there.
2. Cross-jurisdictional expansion. Add a second jurisdiction with a compliance overlay specific to that region. This is a 6–8 week project per jurisdiction, not a one-week project, and shouldn't be rushed.
3. Institutional flow. By year 2, the venue has enough volume to attract institutional desks. Build the institutional onboarding (block trade rails, custom reporting, dedicated relationship management). The institutional share of volume should grow from under-5% in year 1 to 20–35% by end of year 2.
4. Product depth. Year 1 ships a working venue. Year 2 ships a sophisticated one — advanced order types, better analytics for traders, improved position management, educational depth.
5. Brand authority. By year 2, the venue should be producing thought leadership content (research notes, audience studies, market behaviour analysis) that positions the operator as an authority in their vertical. This compounds over years.

The year-2 transition is what determines whether the venue plateaus at year-1 scale or compounds into a durable position. Operators who plan it deliberately out-perform operators who treat year 2 as "more of year 1."

Common mistakes after launch

Five patterns that consistently underperform across the first year. Each is fixable if the operator sees them coming.

Under-investing in operations after launch. Treating the venue as a launched product rather than an operating business. The 4× multiplier mentioned at the top of this post is the gap between operators who actively run their venues and operators who don't.

Listing too many markets too fast. Driven by the intuition that more markets means more revenue. The opposite is true; listing discipline produces the venues that scale.

Aggressive fee changes. Adjusting the fee monthly or based on individual trader complaints. The signal isn't strong enough; the noise overwhelms the signal.

Ignoring resolution operations. Treating resolution as an automated background process that doesn't need attention. Most reputation damage in this category comes from resolution failures that operator attention would have prevented.

Not building the regulator relationship. Waiting until the regulator approaches to start engaging. The relationship has to be built during the calm period; when the cycle tightens, it's too late to start.

Team and hiring patterns through the first year

The team needed to run a venue well evolves across the year, and operators who under-staff at the wrong moments accumulate quiet operational debt that becomes visible only after it's hurt the venue.

Months 0–3 (post-launch). Minimum viable team is one operator-side product person and one part-time market- creation person. Customer support can be outsourced to a generalist provider initially. The protocol provider absorbs infrastructure operations.

Months 4–6. Add a dedicated full-time market-creation person and bring customer support in-house (or to a specialised vendor with prediction-market experience). The support volume by month 4 is enough to justify a named person; outsourcing past this point produces noticeable trader-experience degradation.

Months 7–9. Add a dedicated resolution-operations person or rotation. By month 7 the volume of edge cases per week exceeds what the market-creation person can handle while also creating new markets. The resolution-ops role sits adjacent to support but with deeper authority on specific contract decisions.

Months 10–12. Add a regulatory liaison and a content / communications lead. These are the year-2 critical hires made before year 2 starts. Operators that wait until year 2 to hire these roles spend the first quarter of year 2 playing catch-up on relationships that should have already been in place.

The total team for a venue running well at month 12 is typically 5–7 people on the operator side, plus the protocol provider's infrastructure team that the operator inherits without managing. That's a meaningfully smaller team than the 30–50 person team a from-scratch venue would require, which is part of why the licensed-protocol model has the operational economics it does.

What "running well" actually looks like

A specific definition: at month 12, a venue is running well if it has sustained 15%+ month-over-month WAT growth through the second half of the year, a resolution dispute rate consistently below 0.3%, support volume under 5 tickets per 100 active traders, and the operator team has a documented year-2 plan that the broader business is aligned on.

If you hit those numbers, the venue is on the path to a durable, growing position in its category. If you don't, the diagnostic is which of the running disciplines wasn't operating at the level it needed to be — and the operational improvements at month 12 are the ones that compound through year 2.

The operators who turn into the dominant brands in this category over the next 5 years are the ones who run their venues with this kind of operational discipline from day 31 onward. The launch is necessary but not sufficient. The running is what compounds.