Post World Cup Retention: An Editorial Q&A With Denis Kosinsky, Chief Product Officer at NuxGame, on the Architecture Decisions Taken in August That Shape September

by Luka Dimitrijevic Published on August 11, 2026
Editorial Standards

☆ Editorial Standards

All news content is produced by qualified journalists and analysts under a published editorial code requiring accuracy, source verification, and editorial review prior to publication.

Advertisers and commercial partners have no influence over news coverage.


News editorial policy · Contact us
✓ Fact-Checked

✓ Fact-Checked

Every article undergoes senior editorial review.

Regulatory and legal reporting is cross-referenced against primary sources including official government and regulatory authority records.

Corrections are issued transparently with a visible update notice.


News fact-check policy
⊘ Independence

⊘ Independence

Gamblers Connect is a B2B iGaming media platform.

Editorial decisions, including what to cover, how to cover it, and what to publish, are made independently by our newsroom.

Commercial partners may purchase publication frequency but cannot influence editorial tone, angle, or content.


News independence policy
↗ Commercial Disclosure

↗ Commercial Disclosure

Gamblers Connect is a B2B media platform. We generate revenue through subscriptions, B2B referral partnerships, directory listings, advertising, and media services.

Gamblers Connect is not a licensed gambling operator, affiliate, or player acquisition channel in any jurisdiction.

We do not earn revenue from player activity, wagers, or deposits.


News commercial disclosure · Contact us
The retention stack behind world cup
Key Takeaways
⏱ 19 min read
1
A wide, shallow intake — Optimove Insights' full tournament report, published 5 August, put active bettors at 121% of regional baseline in the US, 136% in Europe and 153% in LATAM, with first time depositors at 179%, 194% and 314%. Average wager per player fell in every region. The tournament delivered breadth, not depth.
2
Retention depends on architecture, not volume — Effective post tournament retention depends on a sequential four layer technical stack consisting of segmentation, lifecycle CRM, cross vertical bridging, and bonus logic operating on live data rather than nightly batches.
3
Strategic cross selling — Cross selling casino products to sportsbook users succeeds best during natural gaps, such as rest days, fixture breaks, or post elimination windows, rather than interrupting players during live matches.
4
Real time responsible gambling — Platform architecture must integrate real time responsible gambling risk profile checks across every retention layer before triggering cross vertical bridges, stake related bonuses, or reactivation messages.

Introduction

The 2026 FIFA World Cup ran 104 matches across roughly six weeks in sixteen host cities across the United States, Canada and Mexico, closing on 19 July with Spain’s win over Argentina. Before the tournament, Eilers & Krejcik projected a US online sportsbook handle base case of 2.82 billion US dollars, with outcomes ranging from roughly 2.32 to 4.33 billion, against an estimated 900 million to 1 billion during the 2022 edition. For sportsbooks operating in the three host markets, the tournament was an acquisition event without recent precedent.

For this Editorial Article, Gamblers Connect spoke with Denis Kosinsky, Chief Product Officer at NuxGame, on why the retention curve after the final is a platform architecture question long before it is a marketing question, and on the four layer stack that decides whether operators can act on the redirecting audience at scale. The Q&A that follows is edited for length and clarity. Selected supporting sources are listed at the foot of this piece.

From Acquisition to Retention

The Post World Cup Retention Challenge

GC: Every operator budgets for the acquisition side of the World Cup. Why do you keep saying the retention side is where the real product problem sits?

Denis: Every operator budgets for the World Cup spike. Almost nobody plans for the Tuesday after, when the same player opens the app again and expects it to remember who they are. That is a product problem before it is a marketing one.

If segmentation, CRM and the bonus engine were built as separate pieces that never really learned to talk to each other, the drop off tends to run quietly for weeks before it registers as a trend. By the time an operator notices the retention curve is sloping the wrong way, the input data that would have let them intervene has already gone stale. That is the uncomfortable part of the timing: the evidence expires faster than the problem does.

Understanding the Post Tournament Audience

What the Data Revealed About Player Behaviour

Why we asked. The post tournament drop off is often treated as inevitable, but the available data presented a more nuanced picture. Rather than assuming that World Cup bettors simply disappeared after the final, this section examines what their behaviour actually indicated and where operators could still influence their next step.

GC: The industry talks about the post tournament drop off almost as a given. Is that actually what the data shows?

Denis: Not in its simplest form, and we can finally argue it from behaviour rather than from a survey. Before the tournament the whole debate ran on stated intentions, which is a bit like judging a restaurant by how hungry people said they were on the way in.

Optimove Insights published a full tournament report on 5 August covering roughly 26 million active bettors a month across the United States, Europe and LATAM, indexed against each region’s own pre tournament baseline. Active bettors averaged 121% of baseline in the US, 136% in Europe and 153% in LATAM. First time depositors averaged 179%, 194% and 314%. I want to be precise about what that proves, though, because the window closes with the final. Those numbers measure the size and shape of the intake. They do not measure what happened to it afterwards, and anyone telling you in August that post tournament retention is solved is reading a scoreboard from the wrong half.

The shape is the part worth acting on. Average wager per player fell in every region while bettor counts rose, so what the tournament handed operators is a wide, shallow cohort, a lot of new accounts staking small amounts. Whether they redirect to a followed player, a rival or a domestic league, or whether they simply stop, is still open. And the one comparable precedent is not encouraging: Optimove’s own read on 2022 was that US World Cup engagement was real but short lived, giving way to local sports cycles with little lasting retention effect. So the retention question is not whether the audience showed up. It is whether the operator can direct them.

The Four Layer Retention Stack

Building Retention Around Connected Infrastructure

Why we asked. With acquisition complete, operators needed a framework capable of turning live player behaviour into targeted retention activity. Denis outlined a four layer approach built around segmentation, lifecycle CRM, cross vertical engagement and bonus logic, with each layer dependent on live data from the previous one.

GC: You describe retention as a four layer stack. What are the layers, and why does the order matter?

Denis: The four layers, in order, are segmentation, lifecycle CRM, cross vertical bridging, and bonus logic. Each one reads live data from the layer below it, which gives the stack the same property as a ladder: skipping a rung does not get you up faster. You cannot personalise a CRM flow off a cohort you have not segmented, and you cannot personalise a bonus off a CRM flow that runs on a nightly batch.

Segmentation groups accounts into cohorts based on live behaviour during the tournament. Lifecycle CRM executes the retention programme automatically off that tag layer, replacing fixed date campaigns with behaviour triggered flows. Cross vertical bridging introduces casino product to appropriate sportsbook cohorts at specific moments where the introduction will not damage trust. Bonus logic fires personalised offers off tournament behaviour, rather than defaulting to a blanket reload.

Where those four layers were assembled as separate systems, the retention plan shows gaps in the same places every year. Where they were built to run as one, the plan holds across the tournament and through the transition into the domestic season.

Segmentation

Identifying Which Players Were Worth Retaining

Why we asked. Not every World Cup bettor represented the same long term opportunity. The most valuable segmentation work happened while the tournament was still underway, using live behavioural signals to distinguish temporary event driven activity from broader betting engagement.

GC: On the segmentation piece, you kept emphasising it has to happen during the group stage. With the tournament finished, what does that mean for operators now?

Denis: It means the highest value version of that work has already closed. By the time the fixture list dries up, the input signal is gone. Behaviour during the group stage and the knockout rounds is a stronger predictor of long term value than any signup form. If you waited until the final whistle to run the segmentation model, the cohort is already fixed. You have descriptive data but nothing to act on, and that is where a lot of teams are sitting this week.

The four signals that carry most of the predictive weight in this window are national team concentration of activitywhether the account touches any cross vertical products at allconsistency of engagement across match days regardless of result, and device and channel behaviour. Those signals need to recompute continuously as data lands, not batch nightly. A tag layer applied once at registration and updated overnight is functionally too slow at the pace of a knockout tournament.

There is one piece of good news for anyone reading this in August. Three weeks of dormancy is a far easier reactivation problem than three months. The data has gone stale, not radioactive. Score the cohort now, on what it did in June and July, and the domestic season gives you several clean re entry points to use it on.

Real Time CRM Architecture

Why Nightly Data Exports Became a Bottleneck

Why we asked. Behaviour driven retention depended on operators being able to act while the behaviour was still relevant. This made the connection between the sportsbook, CRM and bonus engine a critical part of the retention architecture.

GC: You have been direct that a nightly CRM export is the specific bottleneck. Explain what that means in practice.

Denis: Most legacy stacks pull a nightly export from the sportsbook engine into the CRM, and another nightly export into the bonus engine. That cadence was fine when the calendar drove the campaign. It is not fine when player behaviour drives the campaign. Running behaviour triggered retention off a nightly batch is steering a car by yesterday’s traffic report.

If a player’s national team goes out on a Tuesday afternoon, the retention flow that follows them into the next competition needs to fire that same evening, not thirty six hours later after the batch finishes. By then a competing operator has probably reached them. The specific architectural requirement is a live connection between the CRM engine, the sportsbook engine, and the bonus engine, with no nightly export in between.

That one item explains most of the difference between the operators who used the weeks after the final and the operators who watched them pass.

Cross Vertical Engagement

Moving From Sportsbook to Casino at the Right Moment

Why we asked. Cross selling offered another route to extending the value of World Cup acquired players, but the timing of that transition was crucial. Rather than interrupting players during their highest engagement moments, operators needed to identify natural bridge opportunities.

GC: On cross selling from sportsbook to casino, where do most operators get the timing wrong?

Denis: They push casino during a live match. That is the moment when the player is most engaged with the sportsbook product, so it feels like the moment to interrupt them with an offer. In practice it lands badly and costs trust. It is the dessert menu arriving while the main course is still in the air. The player reads the cross sell as pressure.

The design that works is to shift from a single blanket cross sell campaign to two or three specific bridge moments, each mapped to a moment when the introduction reads as helpful rather than intrusive. The three that matter most in a post tournament window are: the immediate window after national team elimination, when sportsbook activity naturally cools; the scheduled break days between fixtures during the knockout rounds; and the post tournament non match days, when the domestic league schedule has not yet fully picked up.

That third window is closing as we speak. The Community Shield is on 16 August, La Liga returns in mid August, the Premier League opens on 21 August, the Champions League draw is on 27 August and the league phase runs from 8 to 10 September. Anyone still planning a cross sell bridge for “the quiet period” has about a week of it left.

Personalised Bonus Strategy

Replacing Generic Reloads With Behaviour Based Offers

Why we asked. Once player behaviour had been captured and segmented, bonus logic provided another opportunity to personalise the retention journey. The focus shifted from generic reload incentives toward offers linked directly to tournament activity.

GC: You have argued that generic reload bonuses are the weakest bonus format for this window. What replaces them?

Denis: Offers that respond to what the player actually did during the tournament. A generic reload is a birthday card with the name left blank, technically a gift, and everyone can tell. Optimove’s post tournament guidance points the same way: prioritise second deposits, repeat behaviour, activity outside World Cup matches, and sustained betting without repeated incentives. That last phrase is the one to sit with. An account that only stays because of incentives is not retained, it is rented.

The three bonus triggers that carry most of the personalised offer traffic in a working retention framework are: a free bet tied to a followed playera boosted price on the next fixture involving a rival team, and a casino credit unlocked when a followed team is eliminated. Each fires off account behaviour, and each is capped at a per player affordability limit set by the responsible gambling layer. Personalisation does not remove the RG cap. It makes the cap more relevant.

One caveat that matters far more in 2026 than it did in 2022: that third trigger is a problem in Great Britain. UK remote licensees have had to collect direct marketing preferences at product and channel level since 1 May 2025, and since 19 January 2026 a single incentive must not include more than one gambling product category, meaning betting, casino, bingo or lottery. The Commission has been explicit that the qualifying activity and the reward have to sit in the same category, and that this applies whether or not a stake on the second product is required. There is an exception for bonus funds usable across all licensed product categories without operator imposed restriction. The customer may choose the category themselves, provided the operator has not narrowed the list. Brazil goes further and prohibits sign up bonuses outright. Whether a specific offer clears any of this depends on how it is constructed, so it belongs with compliance rather than on a product roadmap, but the design principle is safe enough. Where the bonus lever is narrowed, the same segmentation output has to drive non promotional personalisation instead: which competition appears first in the app, which markets are surfaced, which fixtures trigger a notification. The segmentation earns its keep either way.

Responsible Gambling

Making Responsible Gambling Part of the Architecture

Why we asked. Personalisation and retention strategies could not operate independently of responsible gambling requirements. With behavioural data powering both retention and risk detection, RG needed to be embedded across every stage of the framework.

GC: How does responsible gambling sit inside this framework? Because operators can build all four layers and still get RG wrong.

Denis: Responsible gambling is not a separate layer alongside the retention stack. It is inside every layer. The behavioural signals used for retention segmentation are the same signals used for at risk detection under most tier one licensing regimes. If a platform is retention aware but not RG aware, it is not fit for a regulated market.

The concrete requirement is an RG risk profile check at three specific points: before any cross vertical bridge fires from sportsbook to casino, before any personalised bonus with a stake related trigger is offered, and before any reactivation message is sent to an account that has been previously flagged for cool off. Brazil, for one, requires all marketing, bonus offers and push notifications to self excluded bettors to be suppressed for the duration of the exclusion. If the platform architecture cannot make those checks in real time, the retention plan is running outside its own safety margin, and that is a compliance risk before it is a commercial one.

I would add one point that this tournament sharpened. The intake was very wide and very shallow, a lot of new accounts staking small amounts, in markets where gambling related household debt is now an active political subject. Aggressive reactivation aimed at that cohort carries a reputational bill larger than the revenue.

The Operating Cadence

Where Retention Plans Broke Down

Why we asked. The effectiveness of the retention framework depended not only on technology but also on when operators reviewed, segmented and acted on their data. Denis identified two specific points in the tournament lifecycle where execution commonly broke down.

GC: Talk us through the pre tournament, in tournament, and post tournament operating cadence. Where did retention plans break this year?

Denis: They broke in two specific windows. The first was at T minus 30 days, the architecture review, because the CRM to sportsbook and CRM to bonus engine live connections are rarely audited before the tournament kicks off. Teams assume they are live because someone said they were live a year ago. The second was at T plus 7, the cohort separation, because operators rushed to reallocate spend from acquisition to retention before segmentation had fully caught up with the final week of the tournament.

Both are avoidable with a documented checklist, and all five items on ours remain relevant today: live segmentation computing on the June and July data you already hold; CRM triggers replacing fixed date calendars; cross sell bridge moments scheduled and gated by RG risk profile; bonus triggers configured off tournament behaviour inside RG affordability rails, where the market permits them; and an engagement layer scoped to carry through to the Champions League league phase in September rather than stopping at the first domestic weekend.

Measuring Retention

Why NuxGame Chose Not to Publish a Retention Forecast

Why we asked. With operators looking for evidence of how World Cup cohorts would perform after the tournament, modelling could provide useful direction. However, Denis drew a distinction between a directional model and measured results, particularly while post tournament data was still developing.

GC: You modelled the retention curve for the 90 days after the tournament under two configurations. You have decided not to publish the numbers. Why?

Denis: Because a modelled number and a measured number look identical the moment either one leaves its context. Put an illustrative figure beside a real one and within a week somebody has it in a board pack as though it were audited.

The model itself is directional, not predictive. It shows the shape of a retention curve under two platform configurations, coordinated against uncoordinated, so our own teams could reason about where the framework changes the slope. It was never a forecast for a specific operator, market or cohort, and it is not a benchmark anyone should measure themselves against, which is the whole problem with publishing it while no measured 2026 curve exists to check it against.

What the measured data supports is narrower and more useful: the intake was real and regionally uneven, and the engagement uplift was much smaller than the intake in every region. The first hard read on what happened after the final arrives with July and August regulatory filings and Q3 operator reporting. Anyone showing you a finished curve this month drew it from assumptions.

The NuxGame Platform

Bringing the Retention Framework Back to the Platform

Why we asked. After examining the operational and technical requirements behind post tournament retention, the conversation returned to NuxGame itself. The final part of the Q&A looks at how its open ecosystem supported the infrastructure required to connect these different layers.

GC: Bringing it back to NuxGame specifically, what is NuxGame actually doing in this space?

Denis: NuxGame is a turnkey iGaming platform built on three products developed since 2018, a casino platform, a sportsbook, and a game aggregator, with a vetted partner ecosystem around them covering CRM, payments and payment orchestration, compliance, and affiliates. The platform is used by operators launching new brands, and by established operators adding a content aggregator to widen library depth without stacking direct supplier integrations.

CRM sits in the partner layer rather than in our core stack, and I would rather be straight about that than blur it. The point of an open ecosystem is that those partners are selected, integrated and running in comparable markets before an operator goes live, so the connections underneath those four layers, meaning CRM to sportsbook, CRM to bonus engine, and real time RG profile checks, are our responsibility to maintain rather than an integration project the operator inherits three weeks before a tournament. Put plainly: the operators who had those connections working before June spent July segmenting. The rest spent July exporting spreadsheets.

The Takeaway for Operator Product Teams

Turning Tournament Data Into Long Term Value

Why we asked. The World Cup may have ended, but the data generated during the tournament remained valuable. We closed the Q&A by asking Denis for the one practical recommendation operator product teams could act on immediately.

GC: Final question, one thing you would tell operator product teams right now.

Denis: Audit the CRM connections. Not the CRM tool, not the marketing team’s calendar. The connections. Confirm the sportsbook engine and the bonus engine are talking to CRM in real time, not overnight. If either connection is a nightly batch, the retention plan is already running with one hand tied. Fix it before the domestic season is fully underway, and the rest of the framework can execute.

The reallocation from acquisition spend to retention spend needs the segmentation input to be clean. If the segmentation is running live, the reallocation can be aggressive without wasting budget. If it is not, the reallocation is a guess, and guesses are expensive in August.

Five Decisions to Lock In Now

Drawing on the expert interview above, Gamblers Connect editorial summarises the five decisions operators should lock in now that the tournament is over. These are the decisions Denis and the NuxGame team have identified as the ones that pay back most reliably in the September to October horizon when the retention curve becomes visible in the numbers.

  1. Live segmentation is running. Behavioural signals compute continuously. Tag layer is readable by CRM and bonus engines in real time. Not a single tag applied at registration and updated overnight. The June and July behaviour is still on file, so score it before it ages further.
  2. CRM triggers have replaced fixed calendars. At risk flows and lapsed player flows are separated. Both fire on behaviour, not on date. The transition flow to the next competition is drafted and gated by the segmentation output.
  3. Cross sell bridge moments are scheduled. Two or three specific moments, none of them inside a live match window. Fresh RG risk assessment before each bridge fires. Cross sell to casino is blocked for cohorts flagged for cool off in the last 30 days, and in the UK requires product level and channel level consent before it can be sent at all.
  4. Bonus triggers are tied to tournament behaviour. Free bets, boosted odds and casino credits fire off account behaviour, not generic schedules. All triggers stay inside published RG affordability limits and inside the promotional rules of each market. Bonus engine syncs back to CRM for follow up.
  5. The transition engagement layer is built. Missions, streaks and loyalty rewards bridge the gap to the domestic league openers and the Champions League. The Premier League opens on 21 August, the Champions League draw is on 27 August and the league phase runs from 8 September. The layer should be scoped to hold across that whole stretch, not just the opening weekend, so the retention curve has time to hold its shape.

Sources
11 sources verified before publication. All sources in this article trace directly to official websites and statements and their data. How we verify sources →
1
Editorial interview with Denis Kosinsky, Chief Product Officer at NuxGame
Gamblers Connect · GC Original Exclusive Interview
"Every operator budgets for the World Cup spike. Almost nobody plans for the Tuesday after."
2
NuxGame official website
NuxGame · Official Body Industry Partner Primary
Turnkey iGaming platform built on three products developed since 2018, a casino platform, a sportsbook, and a game aggregator
https://nuxgame.com/ ↗
3
Denis Kosinsky, Chief Product Officer at NuxGame
LinkedIn · Primary LinkedIn
https://www.linkedin.com/in/denis-kosinsky/ ↗
4
World Cup 2026 Drives Bettor Acquisition Spike Across All Regions, According to Optimove Insights Report
Optimove Insights, distributed via GlobeNewswire · Agency / Wire Proprietary Data Press Release Data / Report
Active bettors averaged 121% of baseline in the US, 136% in Europe and 153% in LATAM. First time depositors averaged 179%, 194% and 314%.
GlobeNewswire ↗
5
Optimove 2026 World Cup Playbook for Sportsbooks, Part 2, Pre World Cup Report
Optimove · Primary Proprietary Data Data / Report
Prioritise second deposits, repeat behaviour, activity outside World Cup matches, and sustained betting without repeated incentives
Optimove ↗
6
World Cup could generate over $4bn US handle
Eilers & Krejcik Gaming, reported by NEXT.io · Secondary Data / Report
US online sportsbook handle base case of 2.82 billion US dollars, with outcomes ranging from roughly 2.32 to 4.33 billion
Next io ↗
7
Improving customer choice on direct marketing, summary of responses and our position
UK Gambling Commission · Official Body Primary Public Filing
Requirement for remote licensees to collect direct marketing preferences at product and channel level
Gambling Commission ↗
8
Socially responsible incentives, what operators need to know (SR Code 5.1.1(3b))
UK Gambling Commission · Official Body Primary
A single incentive must not include more than one gambling product category, and the qualifying activity and the reward have to sit in the same category
Gambling Commission ↗
9
Brazil Bets Law framework, bonus prohibition and PL 2,258/2026
SBC News, reporting on the Brazilian government · Secondary Court / Legal
Brazil prohibits sign up bonuses outright, and requires marketing, bonus offers and push notifications to self excluded bettors to be suppressed for the duration of the exclusion
SBC News ↗
10
Dates for 2026/27 Premier League season confirmed
Premier League · Official Body Primary
The Premier League opens on 21 August
Premier League ↗
11
World Cup 2026 stadiums and host cities
FIFA · Official Body Primary
104 matches across sixteen host cities in the United States, Canada and Mexico
Fifa ↗
Gamblers Connect only publishes regulatory news after cross-referencing primary government and regulatory sources Read our full editorial standards →
Mentioned in this Article