The Person Who Didn't Make It

You spend twenty minutes on a chart. You paste data, try bar, switch to waterfall, go back to bar. You move the color column, adjust the font, remove two series that clutter the story. You land on something that feels right. The numbers are clear, the hierarchy makes sense, the one thing you wanted to say is the first thing the eye finds.

Then you share it. You paste the link in Slack, or drop a screenshot into a slide, or embed it in a report. And in that moment, something happens that is easy to miss: the chart stops being yours.

The person who opens it has none of your twenty minutes. They did not see the waterfall version. They do not know which columns you removed or why. They do not have the sidebar, the controls, the raw data, the context of your Tuesday afternoon. They have a rectangle with shapes and colors and maybe a title, and they will form their opinion about what it means in roughly five seconds.

Those five seconds are the whole game. Everything you did in the twenty minutes before was in service of them.

The maker’s curse

There is a well-studied phenomenon in cognitive science called the curse of knowledge. Once you know something, you cannot unknow it. You lose the ability to simulate what it feels like to not know.

Charts inherit this problem from their creators. You know that the spike in Q3 was caused by the pricing change, so the spike looks obvious to you. It tells a clear story. But the person seeing the chart for the first time does not know about the pricing change. To them, the spike is ambiguous. It could be seasonal. It could be an error. It could be growth. Without context, the data point is a question, not an answer.

This gap between what the maker sees and what the viewer sees is the central problem of data communication. It is not a design problem or a technical problem. It is an empathy problem. And it shows up in specific, predictable ways.

Three things that get lost in transit

The title that made sense to you. “Q3 Revenue by Segment” describes axes, not meaning. It tells the viewer what the chart contains, not what it says. The title that works for the viewer is closer to “Enterprise segment drove 80% of Q3 growth.” That is not a label. That is the takeaway, delivered before the viewer even looks at the bars.

You probably did not write that title, because when you were building the chart, the takeaway was so obvious to you that labeling it felt redundant. This is the curse at work.

The comparison that lives in your head. You are looking at March numbers, but you are comparing them to February in your mind. The viewer does not have February. They have March, alone, with no baseline. A number without a comparison is just a number. It is not high or low or good or bad. It simply exists.

If the comparison matters, it needs to be in the chart. A reference line. A second series. A before and after. The viewer cannot supply context you did not include.

The column you chose not to show. You removed the “Region” breakdown because it was noisy and distracted from the main point. Good instinct. But the viewer, particularly one in the London office, might wonder why Region is absent. They might assume it was never in the data. They might assume you are hiding something.

Omission is an editing decision, and editing decisions carry meaning whether you intend them to or not. Sometimes a footnote is worth it. “Regional breakdown available in the full dataset.” One sentence that preempts the question.

The five-second test

There is a technique from product design called the five-second test. You show someone a screen for five seconds, then take it away and ask what they remember.

It is devastating when applied to charts. Most charts fail it. Not because the data is wrong, but because nothing is visually prioritized. The viewer’s eye bounces between the title, the legend, the axis labels, and the data, and after five seconds they remember none of it clearly.

Charts that pass the five-second test tend to share a few traits:

The title states a conclusion, not a category. Color is used to highlight one thing, not to differentiate twelve things. There is a visual hierarchy: one element dominates, everything else supports it. And there is negative space. Room for the eye to rest between the thing that matters and the things that provide context.

These are not aesthetic preferences. They are communication decisions. They answer the question: when someone who did not make this chart sees it for the first time, what do they see first?

The screenshot problem

There is an uncomfortable truth about how charts actually travel in organizations. They get screenshot. A chart built in a tool, carefully linked to live data with interactive tooltips and drill-down capability, gets captured as a static PNG and pasted into a Google Doc. Or a Slack message. Or a PowerPoint slide three levels of forwarding removed from the original.

By the time it arrives, it has lost its interactivity, its data source, sometimes its resolution, and often its legend (cropped out by whoever took the screenshot). What remains is the visual impression. The shape. The color. The general direction of the bars or lines.

This is not a failure of the people doing the screenshotting. It is the reality of how information moves through organizations. And it means that a chart needs to work as a screenshot. The meaning needs to survive the loss of everything except the image itself.

If your chart requires someone to hover over a tooltip to understand the key number, that number will not survive the screenshot. If the legend is essential to interpreting the colors, and the legend is at the bottom, it might get cropped. If the takeaway depends on reading the axis labels carefully, it will be lost at Slack resolution.

The question is not whether your charts will get screenshotted. They will. The question is whether they still say something when they do.

Two chairs at every desk

There is a writing exercise where you place two chairs at your desk. You sit in one chair to write. Then you stand up, move to the other chair, and read what you just wrote as if someone else wrote it.

The physical act of changing seats is silly. The cognitive shift is not. It forces you to experience your own work without the context you had when you created it. You notice that the second paragraph assumes knowledge from the first. You notice that the clever reference only works if you already know the source. You notice the gap.

Charts would benefit from a version of this. After building the visualization, before sharing it, sit in the other chair. Look at the chart as if you did not make it. What do you see in the first five seconds? What questions would you have? What is missing?

This is not about dumbing things down. A sophisticated audience can handle a complex chart. But even a sophisticated audience arrives without your context, your twenty minutes of iteration, your knowledge of what the data looked like before you cleaned it up. They arrive cold, and they need the chart to warm them up quickly.

The viewer is the point

There is a gravitational pull in data tools toward the maker’s experience. More controls. More options. More ways to configure, customize, and fine-tune. These are good things. They make the twenty minutes more productive.

But the tool’s ultimate purpose is not to serve the person using it. It is to serve the person who eventually sees what was made with it. The ops manager who opens the link. The executive who glances at the embed. The analyst on another team who encounters the chart in a dashboard they did not build.

That person will never see the sidebar. They will never toggle between chart types. They will never know how many columns were in the original dataset. They will see a rectangle, and in about five seconds, they will decide what it means.

Everything else is scaffolding.

Sankey Flow Studio is built for both chairs. Try it free and share something that lands.