The screenshot beats the infographic

What did we find when we scraped our own LinkedIn account?

We scraped 200 LinkedIn posts this quarter across four accounts, ours and the three accounts closest to ours in audience and subject matter. For each account we measured reactions per post, and we ranked the four feeds on production value, meaning the visible cost of making a post look finished. Designed templates, branded color, careful typography.

Our account ranked first of the four on production value. It ranked last on engagement, and the margin was wide. We measured a median of 24 reactions per post on our feed against medians of 88, 100 and 189 for the three peers, which puts the account that invested the most in how its posts look at about a quarter of the engagement of its nearest peer.

We are publishing the number because losing your own comparison is the strongest reason a reader has to believe the conclusion drawn from it. A firm that wins its own benchmark has produced marketing. A firm that loses its own benchmark has produced a measurement, and the measurement says the thing we were proudest of, the polish, was the thing in the way.

The comfortable explanations do not fit the data. Audience size or posting cadence would produce a gap that ignores format. What the scrape shows is a gap that tracks format exactly. The three accounts that outperform us share a posting format we almost never used, and our default format, the designed infographic, appears in none of their top posts.

definitionPosting the screen

Posting the screen is publishing a capture of real work instead of a designed summary of it. The dashboard as it rendered, the sketch as it was drawn, the tool mid-run with the output still loading. It travels because it is evidence. A reader can see that the work exists, and evidence is the one thing a feed full of compositions cannot supply.

Why does a polished infographic lose to a screenshot?

The two formats answer different questions, and only one of them is the question the reader is asking. The reader of a B2B feed wants to know whether the author does the work being described. A screenshot answers that directly, because a live dashboard, a working folder or a tool mid-run is only in that state when somebody built the thing and ran it. An infographic answers whether the author has access to design, and since every company past a certain size has access to design, the answer carries no information.

A second mechanism sits underneath the first. An infographic arrives finished. The conclusion is decided, the evidence arranged, the rough edges removed, and the reader's only available move is to accept it or scroll past it. A screenshot arrives raw and asks the reader for a small act of verification, reading the numbers, noticing the tab names, deciding what the layout implies. That act is engagement in the literal sense, and the formats that invite it collect the reactions that the formats which forbid it never see.

Nobody at our desk chose wrong on purpose, and neither did the teams whose feeds look like ours did. Polish is the correct standard in every other channel a marketing team ships to. On the website, in the deck, in the campaign, production value signals seriousness, and thin design costs deals, which is exactly why we hold our own site to it. The feed selects for something else, and nobody defines that anywhere, so teams carry the standard they know into the one channel where it works against them. The generous reading is the true one. This is a calibration error, and it persists because almost nobody runs the measurement that would expose it.

The scoreboard makes the error easy to keep. Reaction counts are public, but nobody sees the counterfactual, the reach the same insight would have earned as a capture, so a polished feed reads as a modest success rather than a measured loss. It took a peer set and a median to make our loss visible to us.

What do the three winning posts have in common?

The top post on each peer account in our scrape was a dashboard screenshot, a before and after drawn by hand on paper, and a portrait with a single sentence. Three different formats with one shared property. Each is proof of a person doing real work, captured rather than staged. The dashboard existed before the post did. The pen sketch shows thinking at the speed of thinking, with no design budget anywhere in sight. The portrait attaches a face to a position, which is a form of accountability an anonymous graphic cannot carry.

The detail we keep returning to is the format all three peers use heavily and we almost never used, the screenshot of real tooling doing real work. Our delivery work produces those screens daily. Ad batches mid-generation, scoring passes, reject galleries with the rule that killed each item, working folders with the file names showing. We had been spending effort converting that evidence into compositions, and the conversion subtracted the value, because a composition proves design happened while a screen proves work happened.

There is a name for the structure of this mistake. Stage drift is the gap between the AI GTM maturity stage a company reports and the stage its evidence verifies, and it opens without anyone overstating anything, because no standard existed to check the report against. Our feed had the same structure. The reported level, the most finished-looking account in its set, and the verified level, a median of 24 reactions, had come apart, and for the same reason. Nobody had defined what good looks like on a feed, so we optimized what we could see. We spend our working weeks running verification criteria against other companies' go-to-market claims. Pointing the same instinct at our own feed produced this post.

Do AI engines reward the same format the feed rewards?

Yes, and the overlap is measurable, because we ran a separate audit of how AI engines answer the buying questions in our market. Across the 703 citations that audit collected, linkedin.com was the single most cited domain, ahead of every publisher, every analyst and every vendor site, with Reddit second. The cited items were individual posts by individual operators, not company pages. Given the whole web to draw on, the engines reached for the same class of content the feed rewards.

The cited posts share a shape, and the shape is worth being precise about because it can be built on purpose. They are written in the first person by an operator. They make one specific claim. They carry a concrete number or a named mechanism, something an engine can quote and a reader can check. And they take a position rather than surveying the options. None of the LinkedIn posts cited in our audit were infographics.

The same audit measured which questions earn citations at all, and that narrows the target further. ChatGPT cited sources on comparison questions 75% of the time in our run and never cited definition content. Perplexity cited nearly everything. Claude leaned the other way and cited definitions 67% of the time. Comparison was the only band with a meaningful citation rate in all three engines at once, so the highest-value post takes a position on a genuine either-or question a buyer is already asking, the shape of should we hire a GTM engineer or work with an agency. One artifact then does two jobs. It travels the feed the week it ships, and it keeps surfacing in machine answers long after the reactions stop.

The authorship finding deserves its own sentence, because it decides who should be posting. The engines cite operators by name, and the feed reacts to people over pages, so the citable asset a company builds on LinkedIn accrues to a person willing to hold a position in public. A company that wants the citation has to let a human be the author.

The feed carries this much weight in our market partly because search carries so little. Before any of this work began we measured our own AI search visibility at zero, and demand modeling priced the entire vocabulary of our maturity framework at zero search volume. A buyer in this category rarely types a query and finds a firm. A colleague forwards a post, or an engine assembles an answer from posts it trusts, and both of those doors open to the same format.

What should a B2B team change about its LinkedIn content?

The fix costs less than the habit it replaces. Stop making infographics and post the screen. In practice that is five changes.

• Publish captures, not compositions. The dashboard, the working folder, the reject pile, the tool mid-run. When a post needs a design pass, ask what the design pass is covering for.
• Write as the operator, in the first person. Both the feed and the engines select for a person doing the work over a page describing it.
• Make one claim per post and attach a number or a named mechanism to it. A single checkable assertion is the common element of every cited post in our audit.
• Take a position. Comparison questions were the only band cited by all three engines in our run, and a comparison answer needs a stance, held by someone accountable for it.
• Let production value fall. Ours was the highest of the four accounts we scraped, and it bought a median of 24 reactions.

The capture still has to show something, and the strongest thing it can show is a failure. A reject gallery with the rule that killed each item outweighs a highlight reel, and a scorecard with a failing band outweighs a scorecard where everything passed, because a receipt with no failure in it is a portfolio piece, and portfolio pieces are the most discounted format in this category. Our own losing median is in this post for exactly that reason.

One constraint is real and worth naming. A screenshot of client work is client data, and it needs written permission or it stays private, without exception. The workaround is the work you run on your own systems, which every operating team has more of than it publishes. Your own dashboards, your own experiments, your own reject piles are yours to show, and they carry the same evidentiary weight.

We wrote our own kill criterion before changing anything, because a firm that recommends measurement should be seen submitting to it. If posting the screen does not move our median reach within six weeks, the format explanation is wrong and we will publish that result with the same numbers attached. That is the standard we hold any go-to-market claim to in how we work. The claim states its evidence, the evidence carries a date, and the reader gets to watch.

The infographic era of our feed produced the most professional-looking account in its peer set and about a quarter of the engagement of its nearest neighbor. A scrape we ran against ourselves, and lost, is what changed our behavior, and it will do more to change yours than any advice thread, because the numbers will be yours. Run the same scrape on your account against its three closest peers. The result will be uncomfortable in one direction or the other, and either way you will know what to post on Monday.