Skills Gap Analysis: A Template and a Method That Survives Contact With Reality

A skills gap analysis is a structured comparison between the capability a role requires and the capability the people in that role currently demonstrate. The output is not a report. The output is a short, ranked list of gaps that are worth spending money on.

Most attempts fail for the same two reasons: the skill list grows to eighty rows nobody can hold in their head, and the current-level scores are self-reported guesses. Both are fixable. Keep the list between six and twelve skills per role, and back at least the critical rows with evidence rather than opinion.

Copy this into a spreadsheet. One tab per role, one row per skill, one column block per person. Levels run 0–4 so there is a real midpoint and no false precision.

1. Pick one role, not the whole company. A skills gap analysis that covers every role at once gets abandoned. Start with the role where a capability problem is already visible in a business metric — escalation rate, error rate, ramp time.
2. Write the skill list as observable behaviours. "Communication" is not a skill you can score. "Explains a price increase to a customer without conceding a discount" is. Six to twelve rows.
3. Set required levels before you look at anyone. Doing it in the other order quietly bends the target to match whoever is currently in the seat.
4. Measure the current level with evidence. Manager rating is fine for low-criticality rows. For criticality 3 rows, use something harder: a short assessment, a work sample, a call review, a live observation. Anything that produces a number someone else could reproduce.
5. Sort by weighted gap and act on the top three. Everything below the top three is a note for next quarter, not a plan.

Self-assessment is the fastest way to fill the sheet and the least reliable. People who know a topic well tend to score themselves lower than people who have only met it once. If you use self-rating at all, pair it with one objective measure on the rows that matter, and treat a large disagreement between the two as its own finding — it usually means the role expectations were never made explicit.

A short assessment is the cheapest objective measure. Eight to twelve questions per skill, written so each question maps to exactly one skill, gives you a percentage per skill per person. Track confidence alongside correctness: a correct-but-unsure answer is a gap that has not surfaced yet, and a confident-but-wrong answer is the most expensive kind of gap you have.

Once the grid is populated, three patterns tell you what kind of problem you have. A row that is weak across everyone is a content or process problem — the material was never taught properly, or the process itself is unclear. A row that is weak for two or three individuals is a coaching problem. A row that is strong everywhere is a candidate for advanced content, or for turning those people into the people who teach it.

The fourth pattern is the one worth watching: a row where the required level is 3 and everyone sits at 3. That is not success, it is a ceiling. Either the requirement was set too low, or the role has outgrown the description.

For each of the top three weighted gaps, pick exactly one intervention and one re-measurement date. Not a curriculum — one intervention. A 20-minute walkthrough, a rewritten SOP, a short assessment run weekly for a month.

Then re-measure with the same instrument. This is the part almost everyone skips, and it is the only part that tells you whether the money worked. A skills gap analysis you run once is a snapshot; one you run twice is a measurement.

Annually for the full grid, quarterly for the rows you are actively working on. Anything more frequent than quarterly for the whole grid becomes administrative overhead, and anything less frequent than annually means you are planning against a picture of a team that no longer exists.

Continuous assessment changes this calculation. If short assessments run in the background every week, the current-level column stops being a survey you administer and becomes a number that is simply there when you open the sheet.

One role (support agent), six skills, one person. The weighted-gap column does the prioritising for you: refund handling outranks everything else, even though the raw gap on data classification is the same size — because the business hurts more when refunds go wrong.