Why your board can't read your Level 3 — the cognitive case for summarising

You email the steering committee your 3,000-line control schedule, immaculate and fully linked, and get silence back. It isn't that they didn't try. It's that a logic-linked Level 3 Gantt is, for a reader, effectively unreadable — and no amount of scheduling rigour fixes a communication problem. Here's the case for why summarising, RAG and milestone symbology aren't dumbing down. They're the job.

Let's start with the thing planners hate to admit: the board isn't struggling with your programme because they're not clever enough. They're struggling because you handed them a data structure and called it a chart. A Level 3 schedule is a superb instrument — it's how you run the project. It is a terrible message. Those are two different artefacts, and confusing them is why the one-pager keeps getting rebuilt at 9pm.

The uncomfortable truth is that "just send them the P6 print" has never once worked, on any project, in the history of project controls. Not because the print is wrong. Because reading is not the same as displaying.

Working memory is the hard constraint, not attention span

This isn't a story about busy executives with no patience. It's about how any human reads a picture. Cognitive scientists have a well-worn number for how many distinct objects a person can hold in working memory at once, and it is roughly four — give or take. Not four thousand. Four. A chart communicates when its top-level structure fits inside that budget; it fails when it doesn't.

A well-designed chart respects this by construction. A good bar chart says one thing with five bars. A good line chart has a trend and maybe a threshold. The reader's eye lands, the "so what" arrives in under a second, and everything else is supporting detail available on request. Your Level 3 does the opposite: it presents three thousand equally-weighted objects and asks the reader to do the summarising in their own head, in real time, in a meeting. Nobody can. So they fall back on the only signals available — the tone of your voice and how worried you look.

Signal, noise, and the curse of detail

Here is the paradox at the heart of it. The very completeness that makes a Level 3 trustworthy is what makes it mute. Every activity you add is, from the schedule's point of view, more fidelity — and from the reader's point of view, more noise. Detail and legibility trade off against each other, and past a few dozen bars the trade goes badly.

Planners feel this as a moral hazard. Summarising feels like hiding something, like you're one collapsed WBS node away from spin. So we over-correct: we show everything, on the theory that more truth is more honest. But a message that can't be read isn't honest — it's just unfalsifiable. The board can't challenge what they can't see, which means they can't govern it either. Signal isn't the enemy of rigour. Noise is.

The discipline of a good summary is choosing what to promote. Four or five phases, because that's the budget. The handful of milestones that gate money and go/no-go. One status colour per phase, so "how are we doing" is answered before anyone speaks. And — non-negotiable — the driving path, because a summary that hides what is pushing the end date is decoration, not a schedule.

What a director actually needs to decide

Strip the meeting back to its purpose and a director is there to make a handful of decisions: are we going to finish on time, is any of that in trouble, does anything need my intervention or my money, and can I sign off telling the truth. That's it. None of those questions is answered by activity 1,873's total float. All of them are answered by four phases, a status colour, the next gate, and the driving path.

So read a board schedule the way a board reads any chart. A few categories. A trend or a threshold. A colour that means good-or-bad. One clear "so what." Milestone symbology and RAG stoplights aren't infantilising — they're the visual grammar that lets a non-planner extract a decision in the ten seconds they'll actually give it. That's not talking down to the board. That's respecting how reading works.

This is a presentation problem, so use a presentation tool

None of this is an argument against Level 3. Keep it. Run the project from it. The scheduling engine — critical path, float, logic, calendars, EVM — stays exactly where it is and stays the source of truth. What changes is that you stop shipping the instrument as if it were the message.

This is precisely the job Sketchedule exists for. Import the P6 XER or MSP XML — it parses in your browser, nothing uploaded — and the WBS arrives as an outline. Collapse it: automatic summary bars span each phase, so four bars replace three thousand rows without a single date rekeyed. Colour those phases by RAG from a status column so "how are we doing" is answered before you speak. Promote the gates to milestones with conditional symbology — shape and colour driven by the date or a column value — so a slipped gate looks different at a glance. Highlight the critical phase so the driving path never hides. Then export a branded one-page PDF or PowerPoint, or send a read-only link. Same plan, viewed from an altitude a human can actually read.

Summarising is the harder skill, not the lazy one

There's a quiet snobbery among planners that the "real" work is the detail and the summary is a chore for the graphics team. It's backwards. Building the Level 3 is craft, yes — but deciding which four phases carry the story, which five milestones a director needs, and how to colour the risk so it's honest without being alarmist is judgement. Anyone can show everything. Deciding what to leave out, and standing behind it, is the professional act.

So the next time someone frames summarising as dumbing down, push back. A board can't read your Level 3 for the same reason it can't read a spreadsheet with three thousand rows read aloud — the information is all there, and none of it is communicated. Give them the chart their eyes are built to read, keep the L3 as your source of truth, and you'll find the meeting gets shorter and the decisions get better. That's not a compromise on rigour. That's what rigour is for.

Primavera and P6 are trademarks of Oracle Corporation; Microsoft Project is a trademark of Microsoft Corporation. Sketchedule is an independent product and is not affiliated with, endorsed by or sponsored by Oracle or Microsoft. Figures are illustrative, drawn in Sketchedule; Fig 4 is a faithful redraw of a Sketchedule board-summary view.