What is the PIE framework?

‍

What does each dimension mean?

Potential. How much room for improvement exists. A page already performing well has less headroom than one performing badly, whatever its traffic. Evidence for this comes from funnel analysis and behavioral data showing obvious defects.

Importance. How valuable the affected traffic is. This combines volume with visitor value, so a low-traffic pricing page seen by qualified buyers can outrank a high-traffic blog post. This dimension is PIE's distinguishing feature.

Ease. How little effort the test takes to implement and run, including political difficulty. A homepage test requiring approval from three stakeholders is not easy regardless of the build.

How does PIE differ from ICE?

ICE is the framework PIE gets compared against. The two rank similarly and emphasize different things.

The substantive difference is what the second dimension captures. PIE weights traffic value explicitly, which suits deciding where to focus. ICE weights evidence strength, which suits deciding which hypothesis deserves the traffic.

PIE's gap is that it has no confidence dimension, so a well-evidenced hypothesis and a guess about the same page score identically. ICE's gap is that impact estimates frequently ignore traffic value. Some teams use PIE to select the page and ICE to select the test on it, which covers both.

Neither is a diagnostic. The observations that become the ideas both frameworks rank come from a CRO audit or a page review such as the LIFT model.

Key takeaway: PIE prioritizes surfaces, ICE prioritizes hypotheses.

Where does PIE work best?

Choosing where to start on a large site. When the question is which of forty templates deserves attention, PIE's importance dimension resolves it faster than impact estimates.

Communicating with stakeholders, since potential and importance are intuitive to people outside the CRO team.

Portfolio-level planning, meaning quarterly decisions about which areas of the site get attention.

Where it works less well:

Comparing ideas on one page, where every idea shares the same importance score and the framework collapses to potential and ease.

Any use where scores are assigned without defined scales, which is the failure mode both frameworks share. Write down what each number means before scoring, or the output is a ranked list of preferences.

Related terms

ICE score · LIFT model · CRO audit · Funnel analysis · Landing page optimization

Service: Conversion Rate Optimization.

FAQ

Should you use PIE or ICE?

PIE when choosing which page or area to work on, ICE when choosing which hypothesis to test. Using PIE for the surface and ICE for the idea on it covers both decisions without much extra effort.

Who created the PIE framework?

It was popularized by WiderFunnel as a way to prioritize testing opportunities across a site.

What is a good PIE score?

There is no absolute threshold. An average is meaningful only against the other rows in the same list, scored by the same people in one sitting. A score from last quarter compared against one from this quarter measures scorer drift as much as it measures opportunity.