The data was never the problem. The decision was.

In short: data-driven marketing is not about dashboards. It is about replacing one opinion with one piece of evidence, then acting before the week ends. I built SessionInsight so that any business — one site, one person, no analyst — can see what visitors do, understand why they leave, and fix it. That is the entire vision.

Why I built SessionInsight

I have spent twenty years measuring marketing. Agencies first, then my own consultancy, SEINSIGHTS, working with small and mid-sized businesses — the ones with a real website, real traffic and no data department. For most of that time, I was the person who read the numbers and explained them to everyone else.

Two scenes repeated themselves everywhere I went.

The first: the report. Analytics installed years ago, a dashboard built with care, a monthly email nobody opened. When it did get opened, someone said “interesting”, and the meeting moved on. That was never laziness. It was a report that asked no question. It answered “how many”, when the room was stuck on “why”, and it arrived weeks after the decision had to be made anyway.

The second scene: the wall. A business owner with three hundred visits a month and a form that does not convert, being told that the serious solutions are enterprise analytics, a tag manager, a consent platform and somebody who knows how to wire them together. Everything about that advice was true, and useless. A four-person company does not need a data platform. It needs to see the twelve sessions where somebody typed three lines into a contact form and then closed the tab.

That is the gap I kept standing in. Numbers tell you where people leave. Only the sessions tell you why. Heatmaps tell you what nobody even looked at. Enterprise suites sell all three, wrapped in the complexity of a suite, priced for the companies that can dedicate a person to them.

So I built SessionInsight: one script in the head of the page, three views of the same visitor — what happened, why it happened, what got ignored — hosted in France and Switzerland, with a free plan that is actually usable rather than a countdown. Not to compete with the big platforms, which do a different job. To give small sites the part most tools skip: the why.

What I believe about data-driven marketing

These are the principles I work by, in the order they usually matter. They are opinions — mine, argued in public, and the reason I built a tool that looks the way it does.

1. Evidence should end arguments, not start them.

   Most disagreements about a website are not disagreements about facts — they are disagreements about which of us gets to be right. That is expensive. When two people disagree about a page, the answer is usually one hour away: ten replays, one heatmap, one goal. If a debate survives a second meeting, nobody has gone to look.

2. A number without a question is decoration.

   Before adding a metric, write the decision it could change. If there is no decision, the metric is furniture — it fills a dashboard, it costs attention, and it invites the false comfort of having measured something. A small, honest set of numbers beats a complete set nobody reads.

3. Measure less, decide more.

   You do not need to instrument everything to run a business on data. Pick one goal per page type — a purchase, a sign-up, a request — and let everything else be context. Three numbers you act on weekly will beat thirty you never open, every time.

4. The “why” is never in the chart.

   Charts are the map, sessions are the territory. Use the numbers to find the leak — the page, the device, the source — then watch ten to twenty people stop at it. In my experience the cause is banal and specific: a field that rejects a postcode, a cost that appears one step too late, a button that looks decorative. No amount of aggregation will tell you that.

5. Opinions are welcome. They just have to lose to evidence.

   I am not against intuition; intuition is what writes the hypothesis. But a hypothesis is a question, not an answer. Say it out loud, write it down, look at the sessions and the goal, and let it win or lose. That is the whole discipline, and it costs nothing but a little pride.

6. Every measurement should end in a change — and every change should have a date.

   An insight that ships nothing is a hobby. The other half matters just as much: mark the date on the chart when you change something, or three months later nobody will know whether the trend is your work or the season. One line of annotation saves a quarter of guessing.

7. Small sites deserve clarity, not a smaller enterprise suite.

   The usual answer to a small budget is a stripped-down version of what large companies use: the same vocabulary, the same setup, a fraction of the value. I think that is backwards. A small site needs fewer features and faster answers — one script, no training, a price that is not a project.

8. Privacy is a design constraint, not a feature.

   I ran a survey tool long enough to be uncomfortable with how much personal data people collect by accident. The reasonable position is not “collect first, ask later”: it is to collect what the question requires. That is why SessionInsight has a cookie-less lightweight mode, masks form inputs by default, and lets you turn off IP storage. What you are legally required to do depends on where your visitors are — check with your DPO, not with a marketing page.

9. Trust is built by showing your work.

   Expertise is not a list of credentials; it is the ability to say what a tool cannot do. SessionInsight is not a data warehouse, it does not classify rage clicks for you, and it will not replace an analyst. Knowing the edges of your own instruments is the difference between data and vanity.

How I work with data

Twenty years of audits compress into a short loop. The interesting part is not the size of the measurement, it is the order of the questions: each one is chosen so that the answer makes the next one cheap.

If you want the long version of that loop — including how to read a drop-off without fooling yourself — I wrote it out in why visitors leave your website. The vocabulary I use per page type is in bounce rate vs exit rate, and the habit that keeps it honest is reading a heatmap without over-interpreting it.

What this means for SessionInsight

The product is the argument made concrete. One script gives you the three views I describe above:

• Analytics for the what: sessions, pageviews, events, sources, geography, goals.
• Session replays for the why: the real sequence of scrolls, taps and clicks, with form inputs masked, and the page journey of each visit.
• Heatmaps for what gets ignored: where people click, and how far they ever scroll.

Around that: goals you define yourself, realtime, email reports, shared dashboards, outbound click tracking, and a cookie-less lightweight mode for when you only need the basics. Free covers one site with 50,000 pageviews a month, ten heatmaps, ten goals and thirty days of history — enough to run the whole method on a real business site. Growth is €9 a month for fifteen sites, a million pageviews and a year of retention. The plans are here, and how it works takes two minutes to read.

What it is not. It is not a replacement for GA4 or for a data team, and I will not pretend otherwise on a page about honesty. There is no event log to reconstruct a session by hand, and no automatic rage-click detection: when something looks like a pattern, you will see it in a replay, because you watched it — the tool does not do the thinking for you. It is also not a free trial. The free plan does not expire; the paid plan starts when the free one stops being enough.

FAQ