Every capability review reaches a slide where the funded portfolio is laid over the map. It is the friendliest page in the pack. Around twenty initiatives already under way, each one placed against the capabilities the review assessed, and a caption saying the organisation is already moving in the right direction. Nobody argues with that slide. Everybody's programme is on it.
It is also the only page in the deck that can answer a question none of the others can, and the answer is not the one the caption gives. The caption reads the marks. The finding is in the gaps between them.
A European composite insurance group ran a capability review across part of its retail business, scored where it stood, and then plotted the whole active portfolio onto the same grid. The count is approximate on purpose. Several labels on that plot bundle two or three workstreams under one programme heading, and splitting them into a precise number would be inventing a precision the source does not have. Around twenty is the honest figure, and the argument does not depend on which side of twenty it falls.
The challenge
The trap is reading the plot as coverage. A mark on a map is a claim of intent. It says a budget line exists and points roughly here. It says nothing about how far the capability underneath it will actually move, and two initiatives aimed at one capability can shift it less than one aimed somewhere else.
An earlier page in the same pack had already read current standing against ambition and found gaps on almost everything. That reading is comfortable and useless. If every capability is a gap, no capability is a priority, and the review has produced a list of everything. The portfolio plot is supposed to be the correction, and on its own it is only a different flavour of the same comfort.
What made it useful was the page after it. The same grid was redrawn with the expected step change from that portfolio, banded rather than scored to a decimal, and the lowest band was under half a point of movement. That is the conversion that turns a plot into a diagnostic. It replaces "we have something on it" with "here is how much it will move", and once you have that, you can subtract. Take the expected end state after everything in flight has landed, set it against what the strategy says the business needs, and see what is left.
Ten capabilities were left. Not ten capabilities with problems, which would have been most of them. Ten with problems that the entire funded portfolio, delivered exactly as planned, would not close.
Which specific initiative sat over which specific capability is not a claim this piece will make, because the extract of the deck we work from carries the labels and not their positions, and inferring placement from the order text falls out of a file is how a case study invents its own evidence. The shape survives the loss anyway, and the shape is what generalises.
Two things in that shape are worth sitting with. The first is that the capability the portfolio was expected to move furthest, the single largest step change anywhere on the grid, still appeared on the shortlist of ten. Being funded, and being funded enough, are different tests, and most portfolio reviews only run the first one. The second is that a capability the strategy had deliberately raised above the default level of ambition set for everything else sat in the lowest movement band. The organisation had told itself this one mattered more than average, and had pointed less than half a point of expected progress at it.
Neither of those is a scheduling accident. They are what happens when the portfolio is assembled bottom-up, from divisions proposing work they can execute, and the capability map is drawn top-down, from a strategy nobody costed against the portfolio.
The approach
A plot like this is drawn by hand, from interviews and a delivery register, and it stops being true the moment the portfolio under it moves. Programme boundaries change quarterly. Scope migrates between initiatives without anyone updating a slide that lives in a close-out pack. So the instrument that produced the most useful page in the review is also the least durable thing in it.
That is a solvable problem, and solving it is unglamorous. The join has three columns: the initiative, the capability it claims, and the measurable process the capability actually shows up in. Two of the three already exist in systems. The portfolio sits in a delivery tool with owners, dates and spend attached. The process evidence sits in workflow and policy-administration logs, which is where the operational capabilities can be measured rather than surveyed. Process mining over those logs gives a baseline with a source behind it, refreshed as often as anyone cares to run it, for the capabilities where a survey answer was always going to be the weakest available evidence.
The middle column is the one nobody owns. Which capability does this programme claim, and who agreed? That is a governance artefact, not an analytics one, and it takes an afternoon to create and a standing meeting to keep alive.
This is where a review turns into something we build. The Platform work we do starts with that joining table rather than with the map, because a map redrawn annually by hand cannot govern a portfolio that changes monthly. The division in this group had already picked the right cadence for its own plan, phased into hundred-day increments. A hundred-day delivery rhythm sitting under a once-a-year hand-drawn baseline is a mismatch that will resolve itself the wrong way, with the baseline quietly ignored.
The outcome
What this phase produced was a pilot close-out. A scored baseline, a plotted portfolio, a shortlist of ten, and a set of proposed actions labelled as hypotheses to be validated and priced on a business case before anyone committed to them. Nothing was built here, no initiative was cancelled, and no budget moved. The numbers in this piece are scores, bands and counts, not results.
For each of the ten, the report reached for an outside example of the capability working somewhere else. Five of the ten examples came from outside insurance entirely. That is worth noticing without sneering at it, because it tells you something real: for half the shortlist, the sector had no example to offer, which is usually a sign that the capability is genuinely open rather than merely unfunded here. It also tells you what a set of examples is. Examples establish that a thing is possible somewhere. They say nothing about what it costs in this estate, against these systems, with these interfaces. That distance is the entire gap between an interesting review and a fundable plan, and closing it is what our Consult engagements exist to do.
The durable lesson is smaller than the deck and harder to act on. The portfolio plot is the cheapest diagnostic in any capability review and the one most reliably read backwards. Read the marks and you get reassurance. Read the space between them and you get a list of capabilities with no programme, and therefore no owner, and therefore no one whose quarter goes badly when nothing happens. Those are the ones that are still open next year, and the year after, for exactly the same reason.