Prioritising Experiments with PXL Scores

Most optimisation programmes are limited by traffic, not by ideas. PXL scoring is how Pertento ranks the ideas you have so that the ones worth the traffic run first. You score a draft from its Info tab.

The three dimensions

Each dimension is rated from 1 to 5:

• Potential — how much the change could plausibly move the metric if it works. A checkout change scores higher than a footer change, because more of the decision runs through it.
• Exposure — how many visitors will actually see it. A brilliant test on a page nobody visits produces a result that never reaches significance.
• Level of effort — how cheap the change is to build and ship, scored so that higher is better. A five means trivial to implement. Effort is inverted deliberately: a high total should always mean "do this one", and mixing a low-is-good dimension into the total would break that.

How the score becomes a priority

The three ratings are averaged and rescaled into an importance value from 0 to 10, which is what the draft list sorts on. All three dimensions carry equal weight, so a test cannot buy its way to the top of the list on potential alone while being expensive and barely seen.

A score only exists once all three dimensions are rated. A partly-rated experiment is treated as unrated rather than being scored on the dimensions you happened to fill in.

Where the score shows up

When nothing is broken and nothing is waiting on a decision, your dashboard surfaces your highest-scoring draft as the suggested next experiment, with its score out of 15. That is the point of scoring: on a quiet day the platform can answer "what should I work on" instead of leaving you to re-read the whole backlog.

Scoring honestly

The value of PXL comes from comparing ideas against each other, and it collapses if everything scores well:

• Score against the other drafts in your backlog, not against an absolute ideal.
• Score effort with whoever has to build it, not optimistically.
• Re-score when circumstances change. A test on a page that has since become a landing page for a campaign now has far more exposure than when you scored it.

Notes

• PXL is a prioritisation aid, not a prediction. A high score means the test is worth the traffic, not that it will win.
• All three dimensions are required before a score is stored.
• Scores are visible to everyone on the account, which is what makes them useful for agreeing priorities.