Charts That Sell · Field notes from the pitch room

Choose the chart by the conclusion you're selling, not the data you have

Field note #1 · 6 min read

It's 11:40 on a Friday night. A junior on my team sends me a spreadsheet and a message: "Which chart type would look best for this?"

Wrong question. It's always the wrong question.

Nobody in a client meeting has ever approved a budget because a chart was pretty. They approve it because, for one brief moment, a conclusion became impossible to argue with — and the chart was the thing that made that happen.

Aesthetics are the entry ticket. The argument is the product.

The only question that matters

Here's the part nobody tells you in design courses: in a B2B pitch, a chart is not evidence. It's a prop. You chose the data, you sliced the time range, you decided what to leave out. All of it serves one sentence you need the client to say back to you — out loud, ideally followed by "…so we should do something about it."

So the right question is never about the data you have. It's this: what do I need the client to admit on this page?

My workflow hasn't changed in years. Before I touch a single number, I write the conclusion as the slide title. Only then do I pick the chart whose grammar matches that sentence. And after enough pitches, you notice something: nearly every conclusion you'll ever sell comes in one of four shapes. Each shape has its chart.

1. "Something is changing" → line / area

What you're selling: urgency.

A falling curve needs no adjectives. When a client's organic traffic has slid for six straight months, you can write three paragraphs about why this is concerning — or you can show one line bending toward the floor and let the room go quiet by itself.

The line chart is the only common chart the eye reads left to right, like a sentence. That makes it the native grammar of "this is happening, and it will keep happening." An area chart is the same sentence with the volume turned up: the filled space under the line makes loss feel physical, like watching a reservoir drain.

Rule of thumb: if your conclusion has a deadline in it — "if we wait two more quarters…" — it wants a line.

The tool draws both from the same two columns: a line chart for the series, an area chart for the same series with the volume turned up. And when the change is two periods rather than a continuous run — this year against last year, for a dozen accounts — the slope chart puts every row's direction in one frame.

2. "There is a gap" → bar / hbar

What you're selling: the problem itself.

Gaps only hurt when they stand side by side. "Your acquisition cost is 40% above the industry benchmark" is a statistic. Two bars next to each other — theirs tall, the benchmark short — is an indictment.

Bars are the chart of shame and ambition. When category names get long, tip them horizontal and rank them top to bottom: now you have a league table, and the client can see exactly where they sit in it without you ever sounding like you're attacking them. The chart does the attacking for you.

Rule of thumb: if your conclusion contains the words "compared to," it wants bars.

Both are one paste away: bar chart for short category names, horizontal bar chart when the names are long enough to need a league table.

3. "This is where the money sits" → stacked / donut

What you're selling: reallocation.

Most clients don't have a budget problem. They have a placement problem. A stacked bar showing 70% of spend flowing into the lowest-returning channel does something no memo can: it makes moving money feel obvious, low-risk, almost boring.

That is the entire job of a composition chart — to make reallocation feel inevitable. The moment the client leans forward and says "wait, why is that block so big?", the reallocation has already begun. You're just there to formalize it.

Rule of thumb: if your conclusion is "move the money," show the money as blocks first.

Stacked bar when the mix has to be compared across periods, donut when it's one period and one question. Anything with a second level of grouping inside each block wants a sunburst instead.

4. "Here's where the gap comes from" → waterfall

What you're selling: attribution — which is the pre-sale of your solution.

A waterfall walks the client from last year's number to this year's number, one colored step per cause. Revenue up. Costs up harder. Churn quietly eating the middle. By the final step, the client doesn't just know that there's a gap — they know exactly which section of the bridge collapsed.

And here's the part that matters: one of those steps is the thing your proposal fixes. Once attribution is laid out as a bridge, your solution stops being a pitch and becomes a missing piece. The client walks across it themselves.

Rule of thumb: if your conclusion is "and that's why our fix is the right one," build the bridge.

The waterfall maker builds that bridge from two columns of numbers — where you started, and each cause along the way.

Two ways to give your conclusion away

Telling a trend with pie charts. A pie has no time axis. Four pies in a row, one per quarter, doesn't show change — it asks the client to reconstruct the change in their own head, and they won't. You've handed them the pen. What they write with it is rarely "yes."

Cramming three charts onto one slide to save pages. The client gives each page about three seconds. Three charts is three claims, and three claims is no claim. That deserves its own note — it's field note #2.

The one-sentence test

Before you open any chart tool, write the sentence. Not "Q3 regional performance" — that's a topic, and topics don't close deals. Write "Retention, not acquisition, is where Q3 was won."

If you can't write the sentence yet, no chart type will save you. Opening the tool early just produces prettier confusion.

If you can? Then the rest is mechanics. Paste the table into ChartThat, get all thirteen chart types in seconds, and pick the one whose grammar matches your sentence. The tool's job is speed. Your job is the claim.

— Written between pitches. Field note #1 of Charts That Sell. Next: one slide, one claim.