ANONYMOUS VISITORS
Convert anonymous visitors without identifying them first.
The standard advice is to identify the visitor: enrich the IP, match the device, get the email early. On a considered purchase that mostly produces a worse experience and a data-protection conversation — and it is unnecessary.
You do not need to know who someone is to know what is blocking them. Behaviour inside the current session is enough to form a friction hypothesis and choose a response — no profile, no cross-site identifier, and nothing retained beyond 24 hours.
WHY IDENTITY-FIRST FAILS HERE
Four reasons to stop trying to name the visitor.
Most of them are new
A purchase made once every few years produces first-time visitors. There is no history to look up.
Identity is not a diagnosis
Knowing the company or the segment still does not tell you which of five frictions is live right now.
Early asks cost you the session
An email gate before the buyer has decided anything converts the curious into bounces.
The compliance surface grows
Every identifier you add is another lawful-basis argument. Session behaviour keeps the surface small.
THE USUAL PLAYBOOK
Three ways to spend trust and get nothing.
Email-for-discount gate on arrival
Collects addresses from people who were never going to buy, and irritates the ones who were.
Identity resolution on anonymous traffic
Expensive, patchy, and in the EU a conversation you do not want to have for a marginal gain.
“Welcome back!” to a stranger
The failure mode of profile matching, and the fastest way to look untrustworthy on a €2,000 decision.
WHAT YESWARD DOES INSTEAD
Treat the session as the unit, and forget it afterwards.
Decide from behaviour, not attributes
The hypothesis is built from this visit. A returning visitor is helpful, not required.
Ask for nothing you do not need
The most common response is information, not a form. Level 3 asks one question about the decision, never about the person.
Let consent narrow the action, not break the page
Consent state is an eligibility gate. Less consent means fewer permitted responses and more silence.
Keep the receipt, drop the profile
The decision record survives for measurement; the behavioural profile does not survive the day.
HOW YOU WILL KNOW IT WORKED
One number, one interval, one honest denominator.
Every problem on this page is measured the same way: eligible sessions randomised before treatment, a share permanently withheld, and the difference reported in contribution margin with its confidence interval and its data-quality notes.
95% CI €6.8K–€15.7K · ILLUSTRATIVE DETERMINISTIC DATA · NO LIFT CLAIMED
Bring us the journey that leaks the most.
Two weeks in shadow mode, one journey, one outcome. Every decision recorded, nothing rendered.