A logframe review is really a measurability audit. Test whether the logic holds top to bottom, whether every result can actually be measured, and whether the assumptions survive contact with reality — before a single number is collected.
In short: Review a logframe in three passes — first the vertical logic (do activities produce the outputs, outputs the outcomes, and outcomes the goal?), then the horizontal logic (does every result have an indicator, a baseline, a target, and a means of verification?), and finally the assumptions. The fastest way to sink a report a year from now is an indicator nobody can actually collect, so a good review is mostly a measurability audit.
Read the logframe bottom to top and say the word “if… then” out loud at each step. If we run these activities, then do we actually get these outputs? If we deliver these outputs, then is it reasonable that these outcomes follow? If the outcomes happen, then do they add up to the goal? Most weak logframes break at one specific joint — usually the jump from output to outcome, where “we trained 200 people” is quietly assumed to mean “200 people got better jobs.” Name the broken link rather than smoothing over it.
A common confusion to catch here: an output is what your activity produces (sessions delivered, people trained), while an outcome is the change that results (skills gained, income increased). If a row labelled “outcome” is really counting activity, the whole horizontal row beneath it will measure the wrong thing.
Now read across each row. Every result — goal, outcome, output — needs four things that hold together: a clear indicator, a baseline, a target, and a means of verification. Check each one honestly:
If any of the four is missing or fuzzy, that row is a future reporting problem. Fix it now, while it costs a sentence, not at year-end when it costs a scramble.
The assumptions column is where honest logframes earn trust and lazy ones fall apart. Each assumption is a bet that something outside your control will hold — that participants stay enrolled, that a partner delivers, that the policy environment doesn’t shift. Ask two questions of each: how likely is it to fail, and what happens to the chain above it if it does? An assumption that is both likely to fail and fatal to an outcome is not an assumption — it is a risk that needs a mitigation plan or a redesign.
The slow part of a review is judging, indicator by indicator, whether each one can actually be collected. That is exactly the check you can do systematically. In Sopact Sense you read each indicator against a simple standard — is it specific, does it have a baseline and target, and is its means of verification real — and grade it green, amber, or red. Green indicators are ready to collect; amber ones need tightening; red ones have no realistic way to be measured and would quietly break the report. Reviewing to a fixed standard means two people reach the same verdict, and the weak rows surface in minutes instead of at the funder deadline.
Review each indicator in this logframe: mark it green if it is specific with a baseline, target, and a real means of verification; amber if any of those is weak; red if it cannot realistically be collected. For every amber or red, name the one change that would fix it.
A structured check of a logical framework that tests three things: whether its vertical if–then logic holds from activities to goal, whether every result has a measurable indicator with a baseline, target, and means of verification, and whether its assumptions are realistic. The aim is to catch unmeasurable or illogical rows before data collection begins.
A logic model maps the pathway from inputs and activities to outputs, outcomes, and impact. A logframe (logical framework) is a stricter grid that adds indicators, baselines, targets, means of verification, and assumptions for each level — so it is the measurement and monitoring layer built on top of the logic.
Review it before the program starts, again after the baseline is collected (real data often exposes weak indicators), and at each major reporting point. A logframe that is written once and never revisited is usually measuring last year’s plan.
A weak indicator is one where the wording leaves room for interpretation, there is no baseline to compare against, no realistic source or owner for the data, or it counts activity instead of change. Strong indicators are specific, sourced, and tied to a decision someone will actually make.
Next: How Do You Build a Logframe? · How Do You Build a Logic Model?
ChatGPT, Claude, and Gemini are fine for a quick test — but not for an answer you'll put in front of a funder or board. When it has to hold up, run it in Sopact Sense.
Try it in Sopact →
