Solutions · Managed
We handle the whole process of CRO
Research, hypotheses, build, analysis and the decision. You approve the roadmap and read the results; the programme is ours to run.
What you are buying
How the programme actually runs
Who owns the roadmap, the research-to-decision cycle they run, and what the monthly report has to say.
The engagement
A programme, not a tool subscription
A CRO lead owns your roadmap, runs the cadence and is accountable for the numbers. The platform is included; the work is the point.
- A named CRO lead and a standing cadence
- Research, build, analysis and reporting included
- The roadmap approved by you, executed by us
The process
Research, hypothesis, test, decision
Every experiment starts from evidence (analytics, session data and the surveys we run on your site) and ends with a decision recorded against the hypothesis that produced it.
- Quantitative and qualitative research each cycle
- Hypotheses scored on PXL and worked in order
- A written decision on every experiment, including the losses
Accountability
Numbers you can take to the board
Monthly reporting states the revenue effect, what it does not prove, and what the next cycle will test. Written to survive a finance review, not to flatter the programme.
- Revenue and AOV effect, in your currency
- Losses reported as plainly as the wins
- The next cycle’s plan in the same document
How it works
A repeatable cycle that compounds
From kickoff to the first live experiment in about three weeks. After that it is one cycle a month: identify, build, launch, iterate, each one starting from what the last one proved.
01Identify
Analytics, session data and the surveys we run on your site. Hypotheses scored on PXL and worked in order, not in the order they were suggested.
02Build
Variants built in the editor and QA’d across browsers and devices, including the ones your traffic actually uses rather than the ones we have to hand.
03Launch
Live behind the targeting you approved, cookie-free, monitored from the first session. A sample-ratio alarm catches a broken test before it wastes a week.
04Iterate
A written decision on every experiment, wins shipped and losses recorded against the hypothesis that produced them. Then the next cycle starts.
Cases
What the programme has moved
Three programmes, the number each one moved most, and what the person accountable for it said afterwards.
5×
Return on investment
Sandain
“Very good and well worth testing. We got our investment back in 1.5 months and will probably have a ROAS of 4–5 during the 6-month test period. In addition, the changes you make will remain and generate sales. My strong advice and recommendation is to test them. In the worst case, you will have your money back and gained good knowledge for the future, but more likely you will have increased your sales significantly.”
+14%
Conversion rate
Casall
“Thanks to Pertento we can do more data-driven analysis. Across the tests we run, the uplift comes in at roughly 14 to 20 per cent per experiment against what we had before.”
+17.2%
Add to cart rate
Aim’n
“Verifying hypotheses and making data-driven decisions from test results is hard in an organisation that is constantly moving. What has made our A/B testing work is testing UX and structural changes, and so daring to invest the time in testing before the decision is made.”
Underneath
It is still your account
A managed programme runs inside your own Pertento account rather than inside ours. Everything below is true whether we are running it or you are.
The account is yours
The programme runs inside your own Pertento account rather than inside ours. You can watch every experiment while it runs, not only read about it afterwards.
Read-only access for everyone who wants it
Members see everything and change nothing, so giving the whole marketing team a login costs nothing and risks nothing. Billing is on sessions rather than seats.
The reasoning is in the tool
Every experiment carries the hypothesis behind it and its score on potential, exposure and effort. The backlog is not a document we maintain separately; it is the list you are looking at.
Every change has a name on it
Targeting added, weights changed, started, ended, deployed: recorded against the experiment with who did it and when. A managed programme should be more auditable than an in-house one, not less.
Winners keep working the day they win
A winning variant is deployed to all traffic immediately and stays served by the platform. It reaches your codebase whenever your engineers have room, and the queue keeps flagging it until it does.
Nothing is locked to us
If the engagement ends, the experiments, the history and the results stay where they are. What you would be taking over is a running account, not an export.
Questions
Before you hand it over
Who actually does the work?
A named CRO lead owns your roadmap and runs the cycle, with a research lead, an experiment engineer and an analyst on the programme. You approve the roadmap and read the results; everything between those two points is ours.
What do you need from us to start?
The tag in your head template, access to your analytics, and one person who can approve a roadmap. Nothing else blocks the first cycle, and the tag is the only change to your site.
Do we need our own developers?
Not for client-side experiments: variants are built in the visual editor. Server-side tests are the exception, because there the variant is code in your own application, reviewed and deployed like any other change. Those need your engineers, and we write the specification.
What happens when an experiment loses?
It goes in the report as plainly as a win, recorded against the hypothesis that produced it. A loss that was measured properly tells you something true about your visitors; a loss quietly dropped costs the same traffic and teaches nothing.
Can we take the programme in-house later?
The programme runs in a Pertento account rather than in tooling of our own, so the experiments, the history and every written decision sit in one console. Moving it in-house is a question of who operates that console, not a migration.
How is it priced?
Quoted per programme, with the platform included. The variable is the amount of work rather than the number of people who log in, so it is a conversation rather than a number on the pricing page.