Get Free Assessment Contact

The Experience Optimization Framework

The Experience Optimization Framework is a continuous six-stage loop, Discover, Prioritize, Experiment, Personalize, Learn, Scale, and back to Discover, not a project with a beginning and an end. Scale feeds directly back into Discover, so what’s learned in one cycle compounds into the next instead of the work plateauing after a single push. Every AlexDesigns engagement runs on this loop, whichever stage it currently sits in.

The six stages, in order around the loop

Discover → Prioritize → Experiment → Personalize → Learn → Scale → back to Discover. Two things are true about this loop that aren’t true of a typical process diagram: it never ends, and “Understand” and “Repeat” aren’t separate stages, understanding is what Discover produces, and repetition is simply what the loop does by design.

Discover. Find out what’s actually happening: where a funnel leaks, what a visitor’s real behavior shows, what the evidence says instead of what’s assumed.
Prioritize. Decide which finding deserves the next dollar and the next sprint, weighing impact against effort and risk.
Experiment. Test the change on real traffic instead of shipping it on opinion, so the result is evidence, not a guess.
Personalize. Where a real, meaningful difference between visitors justifies it, show a more relevant experience, never personalization for its own sake.
Learn. Take the result, win, loss, or flat, and turn it into something the next cycle can use.
Scale. Roll out what’s proven, and feed what was learned back into Discover, so the next cycle starts smarter than the last one.

Two AI checkpoints, built into the loop, not bolted on

AI has two specific jobs inside this loop, not a role of its own outside it.

AI premortem on Experiment: model how a test could fail before it ships, not after, so a weak hypothesis gets caught early.
AI post-mortem on Learn: every result, win, loss, or flat, feeds back into Discover through this checkpoint, closing the loop rather than leaving a result stranded.

Accelerate with AI: a ring, not a stage

AI is not a seventh stage. It’s a ring around all six: it speeds up the handoffs between stages, but human judgment still controls every decision that actually matters. AI accelerates; it doesn’t decide.

Why a loop, never a line

A line implies the work finishes. This doesn’t. Scale’s whole purpose is to feed the next Discover, so performance compounds over repeated cycles instead of flattening out after one project. Any description, diagram, or page anywhere that draws these six stages as a straight sequence with an end point is describing something other than this framework.

Where this comes from

For the framework’s own history, tracing back to Alex’s 2015 book, see where the framework comes from.