Every few months a platform changes how it ranks content, and a wave of advice follows explaining what you must now do differently. Post more video. Post less video. External links are penalised. External links are fine again. Founders who organise their marketing around this advice spend their time relearning rules instead of building anything that lasts.
The problem is not that the advice is wrong. Often it is briefly right. The problem is that it depreciates fast, and the effort you spent internalising it does not carry over to the next change. Meanwhile the parts of your marketing that would compound sit untouched.
Why chasing the algorithm never compounds
Think about what you are actually accumulating when you optimise for a ranking system. You are accumulating knowledge of a system you do not control, cannot inspect, and which is deliberately opaque to prevent exactly what you are doing. The half-life of that knowledge is short by design.
Compare that with what accumulates when you optimise for the audience instead: a clearer understanding of who buys from you, a library of posts that still work when republished, an email list, and a record of which messages convert. None of that expires when a ranking model is retrained.
There is a second, quieter cost. Algorithm-chasing content tends to be shaped by the format the platform currently rewards rather than by what you actually have to say. Do that long enough and you lose the thread of your own positioning. The posts perform adequately and mean nothing.
What actually changed when the feed changed
When reach drops after an update, it is worth asking a precise question: did the number of people who could become customers drop, or did the number of impressions drop? Those are not the same thing, and platforms only report the second one.
Reach and revenue often move independently. An update that suppresses link posts can cut your impressions substantially while barely touching sales, because the people who buy from you were already following you and still see your posts. An update that boosts a format can triple impressions with zero commercial effect, because the extra reach lands on an audience with no interest in your product.
If you only measure impressions, both of those look like emergencies or victories when neither may be either. This is the core argument for changing what you look at: not because reach is worthless, but because reach alone cannot tell you whether anything real happened.
The four things you actually control
Strip away everything the platform decides and this is what is left in your hands.
- What you make. Clarity, specificity and usefulness are not subject to ranking updates.
- Where attention lands afterwards. Every post can move someone one step toward a channel you own.
- What you measure. You choose whether your dashboard shows applause or outcomes.
- How consistently you show up, independent of what the feed rewarded this month.
The second one deserves expansion, because it is where most of the durable value is created.
Owned channels are the hedge
An email list, a Discord or Telegram community, an RSS feed, a customer base: these are channels where reaching your audience is not mediated by anyone’s ranking model. Sending an email to a thousand subscribers reaches roughly a thousand inboxes. No update changes that.
The right relationship between social platforms and owned channels is one-directional. Social is where you meet people. Owned channels are where you keep them. Every social post should be doing a small amount of work to move someone across that line, whether that is a link in a profile, an offer at the end of a thread, or simply a reason to subscribe.
Founders often delay building the owned channel because the social numbers are going well. That is precisely backwards. The best time to build the hedge is while the thing you are hedging against is working.
Anchor on outcomes: what to put on your dashboard
The practical version of all this is a change of instrument panel. Instead of impressions and follower count, track a short list of outcomes per post: clicks to an owned destination, signups, trials, purchases and revenue.
Revenue per post is the hardest of these to get and the most valuable to have. Social platforms report attention. Payment processors report money. Nothing in the default setup connects them, so almost no one knows which specific post produced which specific customer. Closing that gap is what seenpaid does: it publishes to 21 networks and reads your Stripe account so each post carries a euro or dollar figure alongside its likes.
You do not need a product to start. Tracked links with UTM parameters plus a note of which post they came from gets you eighty percent of the way, as long as you actually keep the record. The discipline matters more than the tooling.
Building a system that survives updates
A durable content system has four properties. It is format-agnostic, so a change in what the feed favours means reshaping the container rather than rethinking the strategy. It is multi-platform, so a downturn on one network is an inconvenience rather than a crisis. It has a reusable library, so a good post is an asset you can republish rather than a one-time event. And it is measured against outcomes, so you can tell a real change from a cosmetic one.
Multi-platform is the property most people skip because publishing manually to six networks is genuinely tedious. That tedium is the actual reason for single-platform dependence, not a strategic decision anyone made. Remove the tedium and the diversification happens by default.
The library property is similarly underrated. If your last two hundred posts exist only as scattered platform history, every week starts from zero. If they exist as a searchable set with performance attached, every week starts from the best thing you have already written.
A quick audit you can run today
Answer five questions honestly. What share of your audience can you reach without a platform’s permission? If a single network banned you tomorrow, what fraction of your revenue disappears? Can you name the post that produced your most recent customer? How many hours last month went into format experiments versus into understanding your buyer? And when reach dropped, did revenue drop too?
Wherever the answers are uncomfortable, that is the work. It is unglamorous compared with a new posting hack, and it is the only part that is still worth anything in two years.
None of this requires ignoring platforms. Publish widely, use the formats that work, take the reach when it is offered. Just stop letting a system you cannot see set your goals. Whether you assemble that yourself from tracked links and a spreadsheet or use something like seenpaid that does the publishing and the revenue attribution together, the effect is the same: when the number you steer by is revenue per post rather than impressions, the next algorithm update becomes what it should always have been, which is someone else’s news.