Content Engines & Experiments

A $100 million dollar idea sat in a backlog at Microsoft for six months.

An employee on the Bing team suggested a small change to how ad headlines displayed. Management shelved it. Too minor. Not a priority. Then one engineer got curious and ran a quick A/B test.

Revenue jumped 12%. Over $100 million a year. It became the best revenue-generating idea in Bing’s history.

Nobody predicted it. Not the managers. Not the experts. That’s the part I want you to sit with. The people paid to know which ideas would work looked directly at the winner and passed.

Your marketing team is doing the same thing right now. Not because they’re bad at their jobs. Because your organization hasn’t given them a system where testing an idea is cheaper than debating it.

I’ve been having this conversation with clients constantly over the past few months. So today I want to give you the framework I walk them through. It comes down to two words: engines and experiments.

Here’s what’s different about social media in 2026.

Every major platform has quietly converged on the same model. Instagram’s head, Adam Mosseri, has been open about this: the platform shifted from a social graph (showing you content from people you follow) to an interest graph (showing you whatever it predicts you’ll watch). Instagram has confirmed that a quarter of everything users see now comes from accounts they don’t follow. TikTok pioneered it. Reels copied it. YouTube Shorts, LinkedIn, X.

Different logos, same machine.

What that means practically: your follower count no longer caps your reach. Every post competes on merit against everything else in the category. Which sounds terrifying. But it created the single biggest asymmetry in modern marketing.

Content performance follows a power law. Research on platform distribution shows a small fraction of posts capture the overwhelming majority of attention. One study of video platforms found roughly 10% of videos account for nearly 80% of all views.

Translate that into effort terms. You can put 2x the work into a piece of content and get back 100x the result. Sometimes 1,000x. The returns are not proportional to the input. There is no other line item in your budget that behaves this way. Your paid media doesn’t. Your sales team doesn’t. Your trade show booth definitely doesn’t.

Executives understand asymmetric bets. You’ve made them in M&A, in product, in hiring. But when it comes to content, most of you are running a strategy designed to guarantee average outcomes. Safe posts. Approved language. Brand guidelines applied like a straitjacket.

In a power law environment, average is a rounding error.

Now, before you go tell your team to swing for viral moonshots, stop. That’s how companies embarrass themselves. You don’t start with the risky bets.

You start with an engine.

An engine is a content system with a predictable input and a predictable output. You feed ideas in one side. Content your audience actually likes comes out the other. Formats that are proven. A cadence that holds. Performance you can forecast within a reasonable band.

The engine won’t make you famous. That’s not its job. Its job is to do four things:

It gets stakeholders invested. When leadership sees consistent output and steady account growth, social stops being the department nobody defends in budget meetings.

It grows the accounts. Compounding follower and reach growth, week over week. Boring. Essential.

It trains your creative team. Production reps. Your people learn what the demand of real content velocity feels like before you ask them to do anything hard.

It normalizes the cameras. This one gets overlooked. When there’s a lav mic in the conference room every Tuesday, people stop performing. Your subject matter experts get comfortable. The founder stops sweating on camera. That comfort is an asset you cannot buy, only build.

Most companies never get this far. They post sporadically, see nothing, and conclude social doesn’t work for their industry. The engine was never built, so of course it didn’t run.

Here’s my belief: you have to earn your experiments.

Experiments are the moonshots. The formats you haven’t tried. The take that legal will hate. The video that might flop in front of everyone. Most of them will not work, and you should know that going in. Even at Google and Microsoft, companies with world-class testing infrastructure, only 10 to 20 percent of experiments beat the status quo.

That failure rate is exactly why the engine has to come first. The engine gives you the credibility, the audience, and the organizational cover to miss. When 80% of your output is a reliable machine, a failed experiment is a data point. When you have no engine, a failed experiment is proof to the CFO that this whole thing was a waste.

But when an experiment hits? That’s your Bing moment. That’s the post that does the work of a year’s media budget in a week. And you only find it by testing things no one can pre-approve into existence, because if the experts at Microsoft couldn’t spot their own $100M idea, your brand review committee can’t either.

If your company is new to this, my recommendation is one experiment per quarter. That’s it. Four real swings a year. Small enough that no single miss matters. Frequent enough that you’re actually in the game.

I’ll be direct about who this is for.

If you’re running a $20M+ company, you got there by managing risk well. Somewhere along the way, though, risk management in marketing became risk elimination. Every post polished. Every message committee-approved. Every output guaranteed to offend no one and reach no one.

That felt responsible. In the old media environment, it kind of was.

It isn’t anymore. The platforms reward the outlier and ignore everything else. Playing it safe on social isn’t the conservative choice. It’s paying full price for guaranteed mediocrity.

So here’s the assignment. Look at your content operation this week and ask two questions. Do we have an engine, a system that reliably turns ideas into content our buyers respond to? And when was our last real experiment, something that could have actually failed?

If the answer to the first is no, build it. If the answer to the second is “I can’t remember,” you haven’t earned an experiment yet. You’ve just been avoiding one.

Build the engine. Earn the swing. Then take it.

-Jordan

Adam spent 20 minutes looking for a $36 receipt. His finance team sent three Slack messages. Someone made a sticky note.

Ramp would have matched it automatically the moment he swiped. Auto-coded, in-policy, synced. Nobody had to ask Adam for anything.

This is what finance looks like when it runs itself.

Your team can be Adam. Or they can not be Adam.

Don’t be like Adam

P.S. If you want to connect on social media, where I share tips throughout the week, follow me on Linkedin.