The screenshot problem
Every agency deck in this category ends the same way. A grid of posts, a number under each one, and a total at the bottom with an M after it.
Nobody in the room asks what happened next, because the person presenting does not know and the person being presented to does not want to look unsophisticated by asking.
The number is usually real. Impressions are counted by the platform and they are not invented. What is missing is the second half of the sentence.
Reach is an input. It is worth exactly what it causes, and most launch reports stop one step before the only part that matters.
This is why founders discount these numbers on sight, and they are right to. A million views that produced nothing is a screenshot.
Why the second half usually goes missing
Partly because it is genuinely hard. Leaving the app is expensive on X: the reader is mid-scroll, the tab switch costs them their place, and the interested buyer mostly does not click. They read the post, register the name, and look you up on Thursday from a laptop.
By then they arrive as direct traffic or a branded search with nothing tying the session back to a post seen on a phone three days earlier. The launch caused the signup and left no fingerprint on it. We wrote about the measurement problem in full here.
The other reason is less flattering. Reach is the number the agency controls and conversion is the number the client owns, so the report gets built from the half that is easy to produce and easy to look good in.
An agency that never asks what happened in the funnel never has to find out that nothing did.
The question that closes the loop
There is one cheap instrument that works, and it is a free-text field in the signup or demo flow asking how somebody heard about you. Optional, so it costs no conversion. Free text rather than a dropdown, because a dropdown returns the options you already believed in.
It will not be complete. A large share of people skip it, and the ones who answer name whichever touch they remember most recently, which is not the same as the one that did the work.
It is still the only instrument that reaches the buyer who saw the post, did not click, and searched your name two days later.
What it looks like when the loop closes
Codos launched the first virtual Chief AI Officer on 21 September 2026 through the network. The post reached 1.6M people from co-founder Dima Khanarin's account with about 1,800 followers, which is the kind of figure that would normally be the end of the report.
What happened underneath it: 2,031 bookmarks, which for a product sold to CEOs is a buyer saving a vendor to come back to.
The founder's account went from about 1,800 followers to 3,983, so every post since reaches an audience that did not exist the day before the launch.
Then the inbound, which is the part worth arguing about. Two weeks after the post, the founder's account of it was many VCs and many, many customers. We are overwhelmed.
An enterprise product sold to the C-suite, and the thing that filled the inbox was people posting about it.
The honest footnote, which belongs next to the finding rather than in a later paragraph: that is the founder's report, and nobody ran the attribution question underneath it. It says the launch worked and nothing about which channel did the work. The full case study is here.
The ratio nobody publishes
Once you have both halves you can ask the question that actually separates campaigns: how much demand did a million views buy?
It varies by more than most people expect, and the variance is the useful part. A product launch where the thing is live and openable converts differently from a funding announcement, which is news rather than a reason to act.
We measure both across the directory, and funding posts save at 9.4 per ten thousand views against 13.6 for everything else, because a round is something you read and pass on rather than something you come back to.
So the same million views means different things depending on what it was pointed at. That is knowable, and almost nobody publishes it, because publishing it means sometimes reporting a campaign that reached a lot of people and moved nothing.
What to ask for
If you are buying launch distribution, three questions get you most of the way.
- What is our baseline? Demand in the fourteen days before the launch, so the fourteen days after mean something. An agency that does not ask for this before the launch cannot report on it afterwards.
- Which accounts carried it, and what did each one produce? Post-level numbers rather than a campaign total. A total hides the fact that one account did most of the work and six were decoration.
- What did people say when you asked them? The free-text answers, unparsed. If nobody asked, the loop was never closed and the report is the first half of a sentence.
None of that proves revenue, and anyone promising it does is selling you a model rather than a measurement. What it does is move the conversation from how many people saw this to what happened when they did, which is the only version of the question worth paying for.
Questions we get asked
Do impressions matter at all?
Yes, as an input. Reach is the condition for everything downstream and a launch that nobody sees cannot convert. The mistake is treating it as the outcome.
A million views is worth exactly what it causes, which is why a launch report should carry the demand numbers next to the reach numbers rather than instead of them.
How do you measure conversion from a launch when nobody clicks?
With a free-text "how did you hear about us" field in the signup or demo flow, read alongside the change in daily demand against a pre-launch baseline. Neither is complete. Together they catch the buyer who saw the post, did not click, and arrived as a branded search two days later, which is most of them.
Can an agency prove a launch drove revenue?
No, and be careful with anyone who says otherwise. Self-reported attribution is not causation, and a buyer who saw a post, then a press piece, then a message from a colleague will name whichever came last. What an agency can show is the baseline, the change against it, and what people said when asked.
What did the Codos launch actually produce?
The post passed a million views on its first day from a founder account of about 1,800 followers and collected 2,031 bookmarks. Two weeks later the founder's account of it was many VCs and many, many customers, we are overwhelmed, and his account had more than doubled in size.