The segment
Built once, eighteen months ago, and never revisited. It no longer describes anybody.
Hands-on CRM & lifecycle marketing for iGaming
Your customer data already knows what should happen next.
We're the hands-on team that acts on it — inside your platform, shipping every week, and building the tooling most teams wait quarters for. Segmentation, journey orchestration, campaign automation, data quality and reporting, with AI running through all of it so a handful of us carry a department's workload. Built for iGaming — casino and sportsbook — where the second visit is where the money is.
Customer value across the lifecycle
- Segmentation
- Journey orchestration
- Personalisation
- Campaign automation
- AI-built tooling
- Data quality
- Reporting
-
90%+
Top-tier player retention, held
A VIP segmentation, dashboard and retention programme keeping the most valuable players above 90%.
-
~75%
Less hands-on QA time
A 2–4 hour manual campaign-and-template QA pass, down to 30–70 minutes — a person still makes the launch call.
-
~80%
Faster complex email builds
Multilingual templates: 12–20 hours of senior-developer time down to 2–3, then a short human review.
-
15–20×
Faster creative production
A 30-asset banner set — 3 brands, 5 locales, 2 placements — in well under an hour, transcreation included, no translation queue.
Efficiency figures are our own practical estimates on a standard Optimove setup; the retention figure is a live programme result. Happy to walk through the workings on a call.
Most retention problems aren't campaign problems.
The send
The journey looks correct in the builder. The audience arrives at a tenth of its usual size and nobody notices until Monday.
The number
Two dashboards, two answers. So the meeting ends without a decision and the programme stands still.
None of it gets fixed in a strategy document. It gets fixed inside the platform, in the data feeds, and on the send calendar — which is where we work.
What we do
Four areas, and they hold each other up.
Strategy without operations stays on a slide. Operations without data quality ships the wrong thing on time. We cover all four, and you can start at whichever one is currently costing you the most. AI runs through every one of them — drafting and personalising at volume, watching the data, and building the tools — with a person on every decision that reaches a customer.
CRM strategy & lifecycle design
We start from your own behavioural data and work out where value actually sits, then design the segments, tiers and journeys around that. If a programme is already running, we audit what it earns before proposing to change it. What you get isn't a strategy deck that ages on a shared drive — it's the segmentation, the tier definitions and the journey map, specified closely enough to build from the same week. AI does the heavy lifting in the exploration — surfacing the segments and propensity signals in the data faster than anyone working by hand — but the design decisions are made, not generated.
- Segmentation & value tiering
- Lifecycle & journey mapping
- Programme audits
- Channel & contact strategy
Where we start With the segments you already have, and the question of who they actually describe now.
Campaign operations
The weekly grind of getting campaigns out: building them across channels, setting up the promotional and triggered journeys, checking the audience before the send, and keeping the sending reputation clean enough that any of it lands. This is recurring, calendared work, not a project with an end date — the difference between a programme that ships every week and one that ships when somebody remembers. AI drafts and personalises the variants at the volume multi-segment testing actually needs, and the triggered journeys run themselves once built — so the team's time goes on judgement, not production.
- Multi-channel campaign builds
- Promotional & bonus journeys
- Triggered & behavioural messaging
- Deliverability & pre-send QA
Where we start With next week's send calendar, and whether the audiences behind it have been checked.
Data quality & reporting
The part that decides whether the rest works at all. We put validation on the daily imports, watch segment sizes for the drops that break a campaign silently, and agree one definition per metric so the reporting stops being argued with. A segment is only as good as the feed under it, and a report only ends a meeting if everyone in the room agrees what the number means. When a feed looks wrong, we point AI at it and get from “yesterday's audience is off” to a specific cause in an hour instead of an afternoon.
- Feed & batch import validation
- Audience count monitoring
- KPI frameworks & definitions
- BI and insight layers
Where we start With the imports that feed the CRM, and the one metric two teams currently disagree about.
Platform implementation
Getting the CRM properly connected to the systems feeding it, and properly tested before a quarter of planning gets built on top of it. Migrations included — including the unglamorous reconciliation afterwards that proves the move didn't quietly drop a tenth of the file. It's the work nobody notices until it's wrong. And where the stack is missing a piece — a check, a connector, a small internal tool — we build it with AI rather than wait a quarter for it to reach the top of someone's backlog.
- CRM configuration & setup
- Platform & player-data integrations
- Migration and QA
- Team enablement and documentation
Where we start With what's connected to what, and what's been assumed rather than tested.
Selected work
Built, shipped, still in use.
CRM programmes and the internal tools that run them — most built in-house with AI, in weeks rather than quarters, and most still in daily use. Open any one for the detail.
VIP Workspace — the AI daily briefing and today's priority players.
Product screens · sample data
VIP segmentation & retention programme
CRM programme · VIP & retention
VIP segmentation & retention programme
CRM programme · VIP & retention
Problem
A small group of players carries a large share of the revenue, and they're the easiest to lose quietly — to a competitor's offer, or to a lapse nobody caught in time.
Built
A VIP segmentation and retention programme: value-based tiers, the journeys and contact rules around each one, and a real-time dashboard so the people running it can see performance as it moves rather than a month later.
Detail
Tiers rebuilt from behavioural value rather than a static list, lifecycle journeys per tier with responsible-play and eligibility rules sitting above the marketing logic, and a live view of each segment's retention and value so a slip gets attention the week it happens.
Numbers
Retention of the top-tier player base held consistently above 90%.
Outcome
The players who matter most stay, and the team sees them slipping while there's still time to do something about it.
AI-driven campaign QA
Automation · Pre-send QA
AI-driven campaign QA
Automation · Pre-send QA
Problem
Every campaign was checked by hand before it went out — links, dynamic fields, audience size, rendering across clients. It's slow, it's dull, and the one time attention slips is the time a broken merge tag reaches the whole base.
Built
An automated QA pass that runs before every send, with an AI review reading the template and the copy for the mistakes a fixed checklist never catches.
Detail
A pipeline that resolves and tests every link, validates personalisation fields against the data, sizes the audience against history, and hands the rendered output to a model prompted to flag anything off — wrong brand, dead token, a claim that shouldn't ship — before a person gives final sign-off.
Numbers
By hand, campaign QA runs 45–90 minutes and template QA 90–180; with the agent, 10–25 and 20–45. Together a 2.25–4.5 hour pass becomes 30–70 minutes — around 70–75% less hands-on QA time, with a person reviewing the findings and making the launch call.
Outcome
QA went from a slow manual gate to a fast automated one. The person spends their time on judgement instead of clicking every link by hand.
AI HTML email template builder
Internal tool · Email production
AI HTML email template builder
Internal tool · Email production
Problem
Building responsive HTML email by hand is slow and fiddly, and every new campaign or brand meant starting from a fragile copy of the last one and hoping it still rendered.
Built
A builder that turns a brief and a brand into production-ready, responsive HTML email — tested across the major clients — so a template that used to take the best part of a day takes minutes.
Detail
Component-based templates with AI generating and adapting the HTML to brand and content, Liquid branching and all. It preserves the source HTML and typography, validates every language branch, compares the localised compliance blocks exactly, saves to Optimove, then reads the stored version back and checks it again — on top of the usual Outlook, dark-mode and image-blocking traps.
Numbers
A complex multilingual template — Liquid branching, content variations, responsive layouts, localised compliance blocks, Outlook handling, full QA — is 12–20 hours for a senior developer. The builder does it in 2–3, plus a 20–30 minute human review before launch. Roughly 75–85% less build time on the hardest templates.
Outcome
Email production stopped being a bottleneck. New templates and brand variants ship in the time it used to take to brief them.
Multi-brand banner builder
Internal tool · Creative · AI translation
Multi-brand banner builder
Internal tool · Creative · AI translation
Problem
Operators run many brands in many languages, and producing the same banner set across all of them by hand is a slow, error-prone grind — the kind of work a creative team quietly dreads.
Built
A banner builder that generates a full set across brands and sizes from one source, with AI translation built in so each market gets copy in its own language without a separate translation round-trip.
Detail
One design definition rendered across brand skins and ad sizes, AI translating and length-fitting the copy per locale so it still fits the box, with a review step so a person approves the wording before anything goes live.
Numbers
By hand: 20–40 minutes for a banner's first version, then 10–15 per locale after a translation round-trip. With the builder: 1–2 minutes, transcreations included. A 30-asset campaign — 3 brands × 5 locales × 2 placements — drops from ~8–12 hours plus days of waiting to 30–45 minutes: about 15–20× faster, and the designer queue and translation tickets disappear.
Outcome
A campaign's full multi-brand, multi-language creative set is produced in a fraction of the time, without the version-control mess of doing it by hand.
Account management dashboard
Internal tool · Real-time client performance
Account management dashboard
Internal tool · Real-time client performance
Problem
Reporting each client's performance meant pulling numbers by hand before every check-in, and by the time the deck was ready the numbers had already moved.
Built
A dashboard showing each client's CRM and campaign performance in real time, so an account conversation starts from a live view instead of a stale export.
Detail
A live layer over the same modelled metrics the reporting tool uses, split per client, updating as the data lands, with the KPIs an account lead actually gets asked about surfaced first.
Outcome
Client reviews run off one live view. Questions get answered in the meeting rather than in a follow-up email two days later.
Reporting & analytics tool
Reporting · Campaign & player KPIs
Reporting & analytics tool
Reporting · Campaign & player KPIs
Problem
Campaign performance had to be assembled by hand for every review, and two sources rarely agreed on the same number.
Built
A reporting layer with one agreed definition per metric, campaign results and player KPIs in the same view, and enough self-serve that the marketing team stopped queuing for numbers.
Detail
One modelled layer holding a single definition per metric, campaign results and player KPIs joined in the same view, with self-serve filters so a new cut of the numbers stopped being a ticket somebody had to wait on.
Outcome
Reviews spend their time on decisions instead of reconciling two dashboards.
Audience monitoring
Data quality · Pre-send safeguard
Audience monitoring
Data quality · Pre-send safeguard
Problem
Segments can collapse quietly. An audience that normally runs in the thousands returns a handful, the journey fires anyway, and it surfaces weeks later as an odd-looking result.
Built
Daily checks comparing each segment against its own recent history, with thresholds tuned so genuine seasonality doesn't cry wolf.
Detail
A daily job sizing every active segment against its own recent baseline, thresholds set per segment rather than one blanket rule, so ordinary weekly rhythm stays quiet and a real collapse surfaces the same morning it happens.
Outcome
Collapses get flagged the morning they happen, while there's still a decision to make about the send.
Automated data feeds
Data operations · Daily
Automated data feeds
Data operations · Daily
Problem
Daily player data arrived by a chain of manual exports and imports. When a file came in short, the first sign of it was a campaign that had already gone out.
Built
Scheduled feeds between the source platform, the CRM and the warehouse, with row counts and key fields checked on every run and a failure raising an alert rather than a silence.
Detail
Scheduled jobs moving files between the source platform, the CRM and the warehouse, each run reconciled against the last for row count and key fields, with a mismatch paging a person instead of importing a short file and carrying on.
Outcome
Broken data gets caught before the send calendar reaches it, and nobody spends their morning re-running an import by hand.
Marketing planning tool
Internal tool · Planning
Marketing planning tool
Internal tool · Planning
Problem
The campaign calendar lived in a spreadsheet several teams edited at once. Nobody could say with confidence what was going out next week, or who owned it.
Built
A planning tool holding the calendar, the promotional plan, owners and status in one place, structured the way the team actually plans rather than the way a spreadsheet allows.
Detail
A lightweight app over a shared store, with a view per team — calendar, promo plan, owners, status — and a change history, so every edit has a name and a time against it instead of overwriting the last person's.
Outcome
Planning conversations start from the same view. Clashes get spotted before they're built rather than after they're sent.
Telegram sportsbook bot
Channel build · Sportsbook
Telegram sportsbook bot
Channel build · Sportsbook
Problem
Time-sensitive sportsbook messaging doesn't survive the email inbox. By the time a price or an event promotion is read, it's already stale.
Built
A bot delivering odds, event notifications and promotions into a messaging app players already keep open, with opt-in and frequency handled at the source.
Detail
Runs on the messaging platform's bot API, keyed to opt-in stored against the player record, with frequency caps and event triggers so a price or a result fires the moment it's live rather than waiting for the next batch.
Outcome
A channel that arrives in seconds instead of hours, and doesn't compete with a promotions tab for attention.
AI in the work
It changed what a small team can ship. So it runs through everything we do.
This isn't a line on a slide. It's in the working day, every day: writing the checks, drafting and personalising the campaign variants, reading the feed that broke overnight, and building the internal tools that would otherwise have sat in a development queue for two quarters. Every tool listed above was built this way, by the people who needed it, in weeks.
4–20×
Why a small team keeps up Not a headline figure — it's the range across the work above: 4–7× on QA and template builds, 15–20× on creative production. Compound that across a whole programme and a handful of us carry what a full CRM team usually needs — fewer people, less waiting, and the saving passed on rather than padded.
Building tools
Planning tools, dashboards, bots, monitoring, one-off utilities. Anything a marketing team would previously have written a business case for and waited two quarters to receive now gets prototyped, tested and used inside a month.
Campaign production
First drafts and variants across email, push and SMS, at the volume that multi-segment testing actually requires — then edited and approved by a person. Speed at the draft stage, judgement at the approval stage.
Data & QA
Writing the validation checks, interrogating a suspect feed, and getting from "yesterday's audience looks wrong" to a specific cause in an hour instead of an afternoon. This is where it earns the most.
Analysis & reporting
Reading a quarter of campaign results, finding the pattern, and turning it into something a stakeholder will actually finish — without waiting in an analyst queue to learn what happened.
Your stack
We work inside your tools, not alongside them.
No parallel spreadsheets and no shadow process. We've run enterprise lifecycle platforms in regulated gaming environments, and the pattern underneath rarely changes enough to matter — so picking up something unfamiliar is a week, not a quarter.
We name the platforms we're fluent in below — not which client runs what. On a call we'll go through exactly what we've operated and how deep it goes.
- Lifecycle orchestration CRM & campaign platforms — Optimove, Fast Track and similar Segment building, multi-channel journeys, triggers, templates, scheduling.
- Player & customer analytics BI & insight — Qlik, Tableau, Power BI, and custom-built Sizing audiences, reading performance, and holding metric definitions steady.
- AI in the stack Claude, ChatGPT, Copilot, Gemini The leverage behind the tooling and the day-to-day — drafting, checking, building — with a person on every customer-facing decision.
- Core platform data Gaming and sportsbook systems Accounts, wallet, bonusing, betting and game activity as it reaches the CRM.
- Channels & delivery Email, SMS, push, in-app, messaging apps Including the deliverability work that decides whether any of it arrives.
- Data infrastructure Warehouses, feeds and scheduling The plumbing that has to be right before a campaign can be.
How we work
Three steps, in this order.
-
01
Audit
Two weeks inside your data and your platform, scoped and fixed before we start. We read the segmentation, follow the feeds back to their source, sit with the send calendar, and pull on the numbers that don't agree. You get a written assessment: what's running, what's silently broken, what the segmentation is actually describing, and a ranked list of fixes with the reasoning and the rough cost of each attached. That document is yours whatever happens next — including if the honest answer is that you don't need us.
You leave with A written assessment and a ranked, costed list of fixes — yours to keep, or to hand to someone else.
-
02
Build
Segments, journeys, monitoring and reporting get built and tested in your own tools, documented as we go and reviewed with the people who'll run them after us. Where a piece of tooling is missing — a check, a dashboard, a small internal app — we build it rather than route around it. There's no handover event at the end, because none of it ever lived anywhere else.
You leave with Working segments, journeys, monitoring and reporting live in your platform, documented — plus any tooling that was missing.
-
03
Run & optimise
Either we run the campaign operations alongside your team, or we take a monthly cycle: test, read, adjust, report. Retention compounds in both directions, which is why it needs attention weekly rather than a look at the quarterly review — a segment that quietly stops describing anybody costs a little every send until someone notices.
You leave with A programme that gets attention every week, and a monthly read you can actually act on.
Who we work with
iGaming is where we're sharpest.
Casino and sportsbook CRM is what we do most and know deepest — regulated, high-frequency, and unforgiving of a segment that's quietly stopped working.
- iGaming & betting operators
- Casino and sportsbook CRM: player lifecycles, bonus and promotional journeys, reactivation, VIP tiering, churn-risk scoring, and contact rules that respect responsible-play and regulatory constraints.
- Adjacent consumer businesses
- Subscription, apps and marketplaces where retention decides the outcome — onboarding, activation, renewal, win-back, the second purchase. The mechanics are the same; the vocabulary changes, and we pick it up in a week.
- In-house teams needing hands
- Cover for a vacancy, a migration, or a quarter with more roadmap than people. We work to your process, not ours.
Common questions
The things people ask before the first call.
Fair questions, answered the way we'd answer them on a call. If yours isn't here, it's the kind of thing the audit is for.
We already have a CRM team. Where would you fit?
Usually alongside them, not over them. Most in-house teams aren't short on ideas — they're short on the hours to build the monitoring, clear the migration, or cover the quarter where the roadmap outgrew the headcount. We take the piece that's stuck and leave the team running everything else. If what you actually need is a strategy deck and a workshop, we're the wrong call, and we'll tell you so.
Our data is a mess. Is that a dealbreaker?
It's usually the reason to start, not a reason to wait. Almost every programme we pick up has a feed that arrives short sometimes, a segment nobody trusts, or two dashboards that disagree. The audit assumes that going in. We'd far rather find it in week one than have it surface as a campaign that's already gone out to the wrong audience.
Do we have to move to a new platform?
No. We work in whatever you already run. Moving platform is a large, expensive project that fixes far less than the vendor demo suggests, and we'll say so before we'll quote for it. If a migration genuinely is the right move, that's a deliberate decision with its own audit and its own reconciliation — never a default we reach for because something's awkward.
What happens to the tools you build when you leave?
They stay, and they keep working, because they were built in your environment and documented as we went. There's no proprietary layer that stops functioning when the invoice does, and no handover event where the knowledge quietly walks out of the door. That's the whole point of building inside your tools rather than alongside them.
How does an engagement usually start?
With the two-week audit. It's a fixed, bounded piece of work with a written assessment at the end, and it's yours to keep whatever you decide next — including if the answer is that you don't need us. Both sides would rather find that out for the price of a fortnight than the price of a quarter.
How do you handle responsible play and regulatory limits?
As hard limits that sit above the marketing logic, not filters bolted on at the end. Eligibility, contact rules and responsible-play flags decide who can be contacted before any segment does, and a person signs off anything that reaches a customer. It's also the one place we keep AI out of the decision entirely — it can draft and diagnose, but it never approves.
How do you price it?
The audit is a fixed fee. After that it's usually a monthly retainer scoped to a clear set of outcomes, or day-rate cover for a defined gap. We take few clients at a time, so we'd rather quote something we can deliver every week than sell a block of hours nobody will honour. The specifics come on a call, once we know what you're actually trying to move.
Will you name us, or use us as a reference?
Not without asking. No client, platform or employer is named anywhere on this site, and yours won't be either unless you've told us it's fine. The work speaks in categories — what we built and why — not logos. It's the same discretion we'd extend to you.
Contact
Tell us what's not working.
A rough description is enough to start — the platform you're on, what you're trying to move, and what's got in the way so far. Fill in the form and we'll reply with a straight answer about whether we can help.