The Marketing Stack That Apologizes

Open any marketing dashboard right now. I'll wait.

You're looking at a number. The number is presented in a confident sans-serif. It is bolded. It probably has an arrow next to it, green or red, indicating change versus a previous period that someone, somewhere, decided was the relevant comparison. The number does not flinch. The number does not explain itself.

The number is wrong.

Not catastrophically wrong, usually. Wrong in the small, structural ways that all numbers in marketing dashboards are wrong. The pixel didn't fire on Safari. The user opened the email but the image preloader counted that as an open. The form was submitted by a bot. The conversion was attributed to the last touch, but there were eleven touches, and the last one was probably the least important. The cookie expired. The cookie was rejected. The cookie was accepted but the consent banner returned a corrupted string.

I am not telling you anything you don't know. The people who build these dashboards know it. The people who sell them know it. We all know.

We just don't say it.

Imagine, for a moment, that we did. Imagine a dashboard that opened with a small, gray paragraph at the top:

This dashboard reflects approximately 71% of your actual web traffic. iOS Safari users with Intelligent Tracking Prevention enabled (about 19% of mobile sessions) are undercounted. Email opens are inflated by image preloaders, especially Apple Mail's Privacy Protection, which fires opens for ~63% of recipients regardless of behavior. Conversion attribution is last-touch, which is the worst model except for all the other ones. Numbers have been rounded for legibility. We're sorry.

That is the honest version. That paragraph would do more for marketing strategy than the next twelve dashboard updates combined.

I run a small agency. We've adopted, internally, what we call the apologetic footnote. Every report we send to a client includes a section near the top labeled "What this report cannot see." Some examples from recent reports:

We cannot see the people who saw your billboard on the Maine Turnpike, then searched for you a week later from a Wi-Fi network in a Marriott. They appear in this report as "Direct" traffic with no context.

We cannot see the customer who told three friends about your service over dinner. Two of those friends bought from you. The dinner is not in this report.

We cannot see how the people who bounced felt about the bouncing. They might have been annoyed. They might have been satisfied with the answer they got. The bounce rate does not distinguish.

We cannot see whether the increase in branded search this month was caused by something we did, something a competitor did, or simply the calendar.

The footnote does two things. First, it makes the rest of the report read more honestly. The numbers are still there. The graphs are still there. They are just no longer pretending.

Second, and more importantly, it changes the meeting. When clients can see what the report can't see, they stop asking the report to do things it can't do. They start asking us instead. The conversation moves from "what does the dashboard say" to "what do you actually think is going on." That is a much more productive conversation, and it is the one we are paid for.

There's a temptation, in writing this, to position the apology as a humility exercise. Look at us, the agency that admits it doesn't know everything. Trust us, we are honest. This would be marketing. We are too smart for this.

The actual reason to do it is simpler. The footnote prevents you from making expensive mistakes. If your dashboard says paid social is your highest-performing channel, and you do not know that the dashboard counts an open as a click and is using a model that was last calibrated in 2021, you might shift another $40,000 to paid social based on a number that is not real. You will not get $40,000 of result. You will get a story about how the channel "isn't working anymore," and you will probably blame the platform, which will be partly fair and partly not.

If, instead, the dashboard had a small gray sentence at the top that said This number is approximate. We are uncertain about it for the following reasons, you would have started the conversation in the right place. You might have spent the $40,000 on something else. You might have asked a better question.

Uncertainty is information. We have spent twenty years trying to remove it from our reports. The reports have not become more accurate. They have only become more confident, which is a different quality entirely.

I think a lot about a phrase that comes from forecasting research. The best forecasters are not the ones who give you the most precise number. They are the ones who give you a number with the right amount of doubt around it. The number "65%" is not better than the number "between 60 and 70%." It is, in nearly every case, worse. The first is a guess pretending to be a measurement. The second is a guess that knows itself.

We could build dashboards like this. Almost nobody does, because the ones with confident numbers sell better. The people buying them want to feel certain. The people selling them want to sell. The certainty is the product. The truth is a footnote, if it shows up at all.

I will tell you who started doing it well. Weather forecasters. Look at a modern weather app. It does not say "it will rain at 3 PM." It says "60% chance of rain between 2 and 4 PM, accumulation under a tenth of an inch." The forecast is more useful precisely because it admits what it doesn't know. We trust meteorologists more than we trust pollsters, and we trust pollsters more than we trust marketing dashboards, and the gap correlates almost perfectly with how much each profession admits it might be wrong.

I am not asking for a revolution in measurement. The math underneath most marketing analytics is fine. The cookies that survived are still useful. The pixels that fire are still firing. The models, properly used, are still informative.

I am asking for one piece of UI.

A footnote at the top of every dashboard. In the gray, smaller type that humans have learned to read more carefully than the bold, larger type. A short paragraph that lists the things the dashboard cannot see. Updated when the things change. Honest when nothing has changed.

Some marketers will hate this. The ones who hate it will be the ones using the numbers to defend a budget they shouldn't be defending. The footnote will make their job harder. That is fine. That is the point.

The marketers who like it will be the ones who already knew. They have been adding mental footnotes for years, in their heads, every time they open the report. The footnote on the page just lets them point at it instead of explaining it again.

If you build dashboards, build the footnote. If you buy dashboards, ask for the footnote. If you are a CMO sitting in front of a board next quarter, put the footnote in your slide. You will look more credible, not less, because the people who run boards have heard a lot of confident pitches and learned, eventually, what confidence costs.

The most useful number on your dashboard might be the one in parentheses next to all the others.

The one that says: we are working on it.