Why we removed the dashboard chart (and what we learned)

For most of the last year, Purpleturret's main dashboard had a big revenue chart at the top. Beautiful thing. Smooth lines, a hover tooltip, the seven-day average, weekend dimming, a tasteful gradient.

Two weeks ago, we removed it. The dashboard now opens to a single number — yesterday's net revenue — and a smaller weekly grid below. The chart is gone.

This is the story of why, what we tested, what we learned, and what we'd tell other product teams considering a similar move. It's a small story. But it's representative of a larger pattern we keep running into.

What the chart was actually doing

The chart was a hit. Creators loved it. It got mentioned in tweets, screenshotted into pitches, included in the demo we showed to investors. By every superficial measure, it was the most successful piece of the UI.

It was also doing almost nothing.

When we ran a quiet six-week experiment — A/B testing the chart-less dashboard against the chart version — three findings came out:

1. Creators in the no-chart group logged in more, not less. They didn't avoid the dashboard; they came back to it. The frequency-per-day metric went up 11%.
2. Creators in the no-chart group took action on the dashboard more often. "Click a product to see detail," "create a new link," "respond to a refund request" — all went up 7–14%.
3. Self-reported satisfaction did not move. Both groups rated their dashboard experience the same in a quarterly survey.

The chart was beautiful and beloved and slightly making the product worse.

Why a great-looking chart can hurt

A few things going on here, and they're worth naming because we suspect they generalize.

1. The chart is a self-soothing surface. Looking at the chart felt like checking on the business. Buyers loved this feeling. But "checking" is not "doing." The chart absorbed the energy that would otherwise have gone to taking action — creating the next product, optimizing a checkout, answering a buyer email. It was scratching the itch of engagement without producing the value of engagement.

2. The chart over-summarized. Revenue over time is a high-level metric. Most creator decisions aren't made at that level — they're made at the level of "is this product working" or "should I refund this buyer" or "did the new ad creative help." The chart had nothing useful to say about any of those questions. It was the answer to a question creators weren't really asking.

3. The chart anchored emotion. A flat or down day on the chart was a small psychological hit. A flat or down day in a granular product breakdown is just a flat day. Same data, different feeling. The chart concentrated the emotional weight of running a small business into a single up/down line, and that turned out to be a bad emotional design.

What we replaced it with

The new dashboard opens with three pieces of information, in this order:

1. Yesterday's net revenue — one large number. Bold, friendly, no graph. Net of fees and refunds, as it actually lands in the buyer's account.
2. The seven most recent transactions — a small table. Click each to see the buyer, product, link, and any associated events.
3. Three small action prompts — "You have 2 refund requests pending," "Your top product converted at 14% this week," "You haven't shipped anything new in 23 days."

That's it. No chart, no big visualization, no week-over-week sparkles. The page is dramatically less impressive in screenshots. The page is dramatically better at moving creators toward the next thing they should do.

The lesson, less narrow

The chart-removal is one example of a pattern we keep finding inside the product:

The features creators say they love are often not the features that make them money.

Said another way: there's a difference between a feature that delights and a feature that compounds. The chart delighted. It didn't compound. The new action prompts don't delight at all — nobody screenshots them. But they push creators back into the loop of making and selling things, which is the actual job.

We've made a few similar trades since. A "calendar view" of upcoming payouts got replaced with a "next payout in 3 days" pill. A "lifetime revenue" hero number on the analytics page got replaced with "this month vs. last month." A "top-product carousel" got replaced with the single top product with a small "see all" link.

Each one was a downgrade on screenshot quality and an upgrade on what creators actually did with the page.

A note on the resistance

Worth saying: there was real internal resistance to removing the chart. The chart was a thing the team had built carefully and was proud of. Removing it felt like deleting a small accomplishment. This is also a pattern worth naming — teams over-index on features that took effort to build. The cost of building a feature is sunk; the cost of keeping it in the product is recurring (mental space, design constraints, support questions).

The chart took us four weeks to build. The decision to remove it took us six weeks to commit to, even with the data pointing one way. The recurring cost of keeping it would have been forever.

Most products are heavier than they should be because removing things is harder than adding them. Counteracting that requires a real commitment to the boring truth of the data, even when the data is pointing at one of your favorite features.

What you can take from this

Two things, if you build products:

1. Run feature removal experiments, not just feature addition ones. The asymmetry of product work is that we test what we add and don't test what we remove. Half your features are probably not earning their place, but you don't know which half until you try removing them. Pick the feature you're most proud of and the one that's hardest to remove — those are usually where the experiments are most valuable.

2. Distinguish delight from compounding. Build a habit of asking, for each piece of the UI: does this make the user take the next useful action, or just feel good? Both are valuable, but they're different, and most products over-invest in the feel-good half. The compounding half is usually small, plain, and unloved.

We'll probably remove more things this year. We will probably feel bad about removing each of them. The dashboard will look slightly more boring in the screenshot, and creators will get slightly more done. We can live with that trade.

---

Try the new chartless dashboard for yourself. Open Purpleturret — first link in 60 seconds, no chart required.