Assess: Measure the Result
Execute made the change. Assess is where you find out if it worked. This is the step most people skip entirely, and it's the one that turns REPEAT from a to-do list into an actual loop — without it, you're just doing things, not learning from them.
Assess is not the same as Review. Review asks "where do I stand right now, in general?" Assess asks a narrower question: "did the specific thing I executed produce the result I expected?" Review is a snapshot. Assess is a before-and-after comparison against a prediction you made two steps ago.
Don't let it collapse back into Review
This distinction is worth being precise about, because on the surface Review and Assess look similar — both involve looking at numbers. Skip the distinction and Assess quietly turns back into Review: a general check-in instead of a targeted verdict on a specific decision. That's exactly how the loop loses the thing that makes it actually improve over time.
What to measure, and against what
Net worth, savings rate, withdrawal rate, lifestyle impact — the metrics that matter are the ones tied directly to what Plan said would happen. If Plan called for a contribution increase to close an investment-income gap, Assess isn't "how's net worth looking" in the abstract — it's "did investment income move the amount the increase should have produced, given the time elapsed."
This means Assess needs a baseline and a timeframe, both set back in Plan or Execute. Without them, you can't actually assess anything — you can only vibe-check whether things feel like they're going well.
Give it enough time to mean something
| Change executed | Reasonable time before assessing |
|---|---|
| Spending cut | ~1 month |
| Contribution strategy change | At least a couple of quarters |
| Market-exposure / allocation shift | Longer, given ordinary volatility |
Assessing too early doesn't just produce a null result, it produces a misleading one, because normal short-term noise gets misread as the outcome of your decision. Waiting too long slows the whole loop's ability to correct itself. Match the wait to what actually changed.
Separating signal from sequence-of-returns noise
This matters more the closer you are to your target date. A portfolio that underperformed this quarter might mean the plan isn't working — or it might just mean the market had a bad quarter, a completely different problem requiring a completely different response. Conflating them is expensive in both directions: overreacting to normal volatility burns discipline on non-problems, while writing off real underperformance as "just noise" lets an actual issue compound for another full cycle.
A verdict, not a fix
Assess should end with a clear read: the plan worked as expected, it worked but not enough, or it didn't work. What to do about that verdict is Transform's job, not Assess's. The temptation here is to start adjusting mid-assessment — "well if this didn't work, let's also change that" — which blurs measuring and reacting, and is how a clean loop turns into a messy one where you can no longer tell which change produced which result.
What Assess is not
Assess doesn't re-run Review's general inventory — it's targeted, not comprehensive. And it doesn't propose the next move — that's explicitly reserved for Transform, one step later, precisely so the verdict and the response don't get tangled into the same decision made under the same momentary read of the numbers.