Marketing Intelligence Research for an iGaming AI Platform
An AI company building personalisation and LTV prediction tools for casino operators needed to validate their next product milestone before investing in development. We sourced and interviewed five senior iGaming marketing leaders — Heads of Marketing, Heads of Acquisition and CMO-level practitioners — to map how operator marketing teams actually work, where their real pains are, and which AI use cases they would pay for first.
Building AI for operators without knowing how operators actually work
The client was a technology company developing an AI platform for online casino operators — focused on UX personalisation and LTV and churn prediction. The platform had real technical depth and a clear commercial thesis. The problem was the next phase of the product roadmap: Marketing Intelligence.
Marketing Intelligence is not a UI feature — it is an answer to the question “what data-driven decisions do operator marketing teams wish they could make faster?” Building that product without understanding how those teams operate, what tools they actually use, what their manual workflows look like, and where their most expensive blind spots are would mean building for a hypothetical user rather than a real one.
The client needed qualitative signal from practitioners, not survey data or desktop research. They needed to hear from the people who manage traffic, run affiliate programs, calculate CPA, fight bonus abuse and evaluate LTV — in their own language, with enough depth to understand not just what they do but why they do it that way and where they feel most stuck.
Why qualitative research, not a survey
A survey tells you what operators will admit to in a tick-box. A 90-minute expert interview tells you what they actually do, what workarounds they have built, what they are embarrassed about, and what they would change if they could. For product hypothesis validation in a technical vertical like iGaming marketing operations, those two layers of insight are completely different in value.
Five phases. Zero panel shortcuts.
Generic market research panels do not have iGaming marketing practitioners at the seniority level that produces useful product signal. We built the respondent pool from our own network of operators and accessed people who were actively working with CPA, affiliate management, anti-fraud and LTV modelling — not consultants or ex-practitioners describing what the job used to be like.
Expert recruitment
Sourced and engaged 5 respondents from active iGaming operations — Head of Marketing, Head of Acquisition and CMO-level — who were currently working with the systems and workflows the client needed to understand. Each respondent was screened for direct hands-on involvement with traffic channels, affiliate programs and marketing analytics rather than oversight-level familiarity.
Research design
Built a custom interview scenario, discussion guide and respondent interaction flow calibrated to the client’s specific product hypotheses. The research design was not generic qualitative fieldwork — it was structured around the exact questions the product team needed answered to make confident decisions about feature prioritisation and the sequencing of the Marketing Intelligence roadmap.
Expert briefing
Before each interview, respondents received a pre-briefing session to understand the product context and the type of insight the client was seeking. This step significantly improves signal quality: respondents who understand why they are being asked a question provide more precise, more relevant answers than those responding cold to abstract hypotheticals. It also reduces the interview time wasted on orientation.
Deep-dive interviews
Five structured expert interviews of 1–2 hours each, moderated by an iGaming industry specialist. Topics covered: current state of traffic and affiliate channel management, CPA calculation approaches and known distortions, LTV modelling and traffic quality assessment, anti-fraud and bonus abuse detection workflows, existing AI and automation tool adoption, and unmet needs in marketing analytics. Each interview was designed to surface the gap between what operators say they do and what they actually do.
Analysis and reporting
Structured summaries of each interview, followed by cross-interview analysis that identified recurring patterns, contradictions and divergences across respondent profiles. The final output was a full analytical report with conclusions structured around the client’s product hypotheses, plus a feature opportunity list built on stated and observed needs — ready for direct input into the product roadmap.
The six areas where operator marketing teams actually struggle
The interviews were not structured to confirm the client’s assumptions — they were designed to surface what operators actually experience, including things the product team had not anticipated. These six themes emerged consistently across respondents.
Traffic and affiliate channel management
Most operators manage multiple affiliate relationships with heavily manual reconciliation processes. The ability to distinguish between traffic quality at the source level — not just volume — was consistently cited as a gap. CPA models were often based on proxies rather than real LTV signals, creating systematic overpayment for low-quality traffic and underpayment for high-value cohorts.
CPA calculation and LTV assessment
Operators described significant divergence between how CPA is calculated in affiliate agreements versus how it maps to actual player value. LTV modelling exists in most operations but relies on trailing data and human judgment rather than forward-looking prediction. The combination means that acquisition cost decisions are being made with a significant time lag relative to actual player behaviour.
Anti-fraud and bonus abuse
Anti-fraud and bonus abuse detection were described as heavily reactive across almost all respondents. Patterns are identified after loss has occurred, rules are adjusted manually, and the coordination between marketing (who designs bonus mechanics) and risk (who absorbs the consequences) is often weak. Automation in this area was the most consistently identified AI use case with clear willingness to pay.
Manual versus automated workflows
The volume of manual work inside operator marketing functions was higher than the client’s pre-research assumptions. Campaign setup, reporting, affiliate payment reconciliation and retention trigger management all had significant manual overhead even at well-resourced operators. This created specific product opportunities in automation that were not in the original product brief.
AI tool adoption and barriers
Existing AI tool adoption in operator marketing was limited and fragmented. Respondents were aware of general-purpose AI tools but described low trust, poor integration with iGaming-specific data structures, and a lack of context about gambling regulatory environments as the main barriers. An iGaming-native AI product with pre-built understanding of operator data models had a clear positioning advantage.
Differences across team functions
Marketing, affiliate management, BI and risk teams within the same operator had meaningfully different needs, different vocabulary and different priorities when it came to marketing intelligence. A product that spoke only to one function risked being blocked or deprioritised by others during procurement. The research surfaced the decision-making structure and helped the client design a narrative that addressed multiple internal stakeholders simultaneously.
Four structured outputs, ready for immediate use
The research program produced four structured deliverables, each designed to be immediately actionable by a different part of the client’s organisation.
Research Design & Interview Guide
Full interview scenario, respondent interaction flow and structured research framework. Built around the client’s specific product hypotheses rather than a generic qualitative template. Reusable for follow-up research rounds as the product evolves.
Interview Moderation & Transcripts
5 in-depth expert interviews moderated by an iGaming specialist, with full coverage of traffic, affiliates, CPA, LTV, anti-fraud, bonus abuse and AI tool adoption. Structured summaries provided for each interview with key quotes and observations flagged.
Analytical Report
Cross-interview analysis with conclusions structured around the client’s Marketing Intelligence product hypotheses. Covers confirmed assumptions, disconfirmed assumptions, unexpected findings and strategic implications for product positioning. Designed to be presented directly to product and commercial leadership.
Product Roadmap Prioritisation
Feature opportunity list built on stated and observed needs from five senior practitioners. Ranked by frequency of mention, willingness to pay signal and implementation complexity. Provides a defensible, market-evidenced basis for sequencing the Marketing Intelligence roadmap.
Validated product direction before a line of code was written
The most expensive mistake an AI product company can make is building a feature set based on internal assumptions about how operators work — and discovering only at go-to-market that the real workflow is different, the vocabulary is different, and the buying committee has priorities nobody modelled. This research eliminated that risk for a specific product milestone before the development budget was committed.
Beyond hypothesis validation, the research surfaced specific use cases — particularly in anti-fraud automation and affiliate quality scoring — where willingness to pay was demonstrably higher than the client had assumed, and where incumbent solutions had clear weaknesses that the client’s technical architecture could address.
The commercial value of getting this right before development
Product hypothesis validation through expert interviews typically costs 3–5% of a development budget for a single milestone. The cost of building the wrong features and discovering it at launch is typically 40–60% of that same budget in rework, plus the compounding cost of delayed market entry. The ROI on qualitative research at this stage is not abstract — it is one of the clearest risk-adjusted investments a product company can make.
Specific outputs the client could act on immediately:
- Clear picture of how marketing teams make traffic and affiliate channel decisions — including the manual steps nobody admits to in demos
- Pain and limitation map across acquisition, retention, anti-fraud and bonus abuse — with severity scores by function
- Practical insight on CPA calculation approaches, LTV assessment methodology and traffic source ROI logic
- Mapped differences in needs and vocabulary across marketing, affiliate management, BI and risk teams
- Qualified feedback on AI tools for marketing analytics automation — what exists, what is trusted, what is not
- Product hypothesis validation with clear pass, fail and needs-refinement categorisation per hypothesis
- Understanding of the key metrics operators use to evaluate traffic and player quality — essential for product language and integration design
The situations where iGaming expert research changes product decisions
This type of engagement is not useful for everyone at every stage. It is specifically valuable in four scenarios where the cost of being wrong is high and the cost of the research is comparatively low.
Pre-development product validation
When you have a product hypothesis and need practitioner signal before committing engineering resources. Most valuable at feature milestone planning stages where the scope is still flexible and the cost of change is low. Typical engagement: 3–8 expert interviews, analytical report, roadmap input.
Go-to-market positioning research
When you have built a product and need to understand how operators actually evaluate and procure tools in your category. This surfaces the real buying journey, the internal stakeholders involved, the objections that kill deals before they reach your sales team, and the vocabulary your materials need to use.
Competitive intelligence and gap analysis
When you need to understand what existing solutions actually deliver versus what they claim, from the perspective of operators who have bought and used them. Practitioner interviews surface the specific failure modes, workarounds and unmet needs of incumbent tools that competitor analysis and public reviews do not capture.
Partnership and BD due diligence
When you are evaluating a partnership, acquisition or market entry and need rapid expert input on whether the opportunity is real, who the real decision-makers are, and what the typical buying and implementation cycle looks like in a specific vertical or GEO. We can mobilise 3–5 relevant practitioners within days.
More from Uberman
Building something for the iGaming market and need to know if it will actually land?
We design and execute qualitative research with senior iGaming practitioners — product validation, market intelligence, buyer journey mapping. No panels. No proxies. Real operators who are doing the job today.