Methodology
The lab framework, in full
Our homepage says we run SEO like a lab: hypothesis, experiment, measure, scale. This page is the deep-dive on what that actually means in practice for a casino or sportsbook site — how a hypothesis gets written, what an experiment write-up looks like, and how the five steps of our method fit together. No agency buzzwords, just the mechanics, because for YMYL gambling content, “trust us” isn’t good enough.
The framework
Hypothesis → Experiment → Measure
Most agencies pitch a checklist of tactics. We commit to a specific, falsifiable claim before we touch a single page, run the change as a bounded test, and report the real number afterward — win or lose. Here’s what each stage actually requires:
Written before any work starts
A real hypothesis names the page or template, the specific change, the metric it should move, and by how much, in a fixed window. “Improve rankings” isn’t a hypothesis. “Adding FAQ schema to our 40 payment-method pages will lift their average position from #14 to top-10 within 45 days” is.
Scoped, timeboxed, logged
SEO rarely allows a clean A/B split, so we isolate variables the way it’s actually possible to: change one template or cluster at a time, hold everything else steady, log the exact deploy date, and pick a control group of comparable pages left untouched wherever we can.
Judged against the pre-registered number
At the end of the window we report the actual result against the number we committed to upfront — not a reframed metric that happens to look better. If it missed, we say so and diagnose why. If it worked, it gets scaled to the next template; if not, it’s retired and the queue moves on.
Why this matters for gambling content specifically: Google applies YMYL scrutiny to anything touching real money. A pre-registered, falsifiable process is also how we keep ourselves honest on compliance — we can’t quietly redefine success after the fact on a responsible-gambling or licensing-related page.
The five steps, expanded
How the homepage’s five-step method actually runs
Same five steps as the homepage overview — here’s the detail behind each one, and which team role owns it. See our team page for who’s behind each role.
Forensic audit & baseline
Before any hypothesis gets written, we need a true baseline: current rankings per template, crawl logs, cannibalized bonus pages, and where your player intent is actually leaking to competitors. This baseline is what every later hypothesis gets measured against — without it, “improvement” is unverifiable.
Architecture & crawl control
Faceted game filters, state/province page sprawl, and JS-rendered lobbies are where iGaming sites bleed crawl budget. Fixes here are shipped as discrete, logged changes — each one is itself a small experiment (e.g. “canonicalizing filter params will reduce indexed duplicate pages by X% within 30 days”), verified against server logs, not guesses.
Entity & content engine
This is where topical mapping happens: identifying every entity a player or regulator associates with a topic (RTP, licensing body, banking method, wagering terms) and building a hub-and-spoke content cluster that covers all of them, with question-led headings and direct answers up front. It’s the highest-leverage step for YMYL trust because it’s where factual accuracy and depth live.
See the entity categories we map for every cluster
Measure, prove, scale
Every experiment from steps 2–4 gets a retrospective against its pre-registered KPI at the end of its window. Winners get scaled to the next template or market; losers get a documented reason why and are retired. This is also where NDC (new depositing customer) attribution gets modeled back to specific pages, not just traffic.
Proof, not promises
Example experiment write-ups (redacted)
Operator names and exact URLs are redacted under NDA, but the structure below is exactly what you’d see in your own weekly report — hypothesis, method, and the real result, including the one that underperformed.
Restructuring 40 payment-method pages into a hub-and-spoke cluster with FAQ schema will lift average position from #14 to top-10 within 45 days.
Canonicalizing faceted odds-filter URLs will cut indexed duplicate pages by 60%+ and recover crawl budget for money pages within 30 days.
A digital PR push around a crypto-withdrawal-speed study would earn 8–12 links from fintech publishers within 60 days.
We include the missed experiment on purpose. A methodology that only shows wins isn’t a methodology — it’s a highlight reel. Every engagement gets the full experiment log, not just the ones that worked.
Want your first hypothesis written for free?
The free audit includes one fully-scoped experiment for your site — hypothesis, method, and the metric it should move — so you can see the framework applied before committing to anything.