The number, and why it went up
Advertising has argued about effective frequency for seventy years and the old rule of thumb was about seven exposures. Every team we work with that has measured it honestly in the last few years lands somewhere higher, and one of our clients puts it at roughly fifteen before an average user creates an account.
Two things pushed it up. The feed got far more crowded, so each encounter carries less weight than a full-page advertisement did in 1970. And the encounters themselves got shallower: a clip watched for two seconds while scrolling is a real touchpoint and a weak one.
Treat fifteen as an order of magnitude rather than a target. The useful move is not hitting a number, it is abandoning the assumption that anybody buys on first contact, because almost every distribution plan we see is built as though somebody might.
Different encounters, not repeated ones
This is the part that changes what you do on Monday.
Fifteen impressions of the same advertisement is frequency, and past a low threshold it produces irritation rather than familiarity. Fifteen encounters of different kinds, from different sources, in different registers, produce something else entirely: a sense that the company is simply around, which is what people mean when they say they have heard of you.
| Encounter | What it contributes |
|---|---|
| A founder clip | A face and a point of view. The most memorable single form, and the hardest to scale. |
| A creator covering it | Third-party credibility. Somebody with their own reputation chose to mention you. |
| A podcast appearance | Depth and time. Forty minutes of someone's attention is worth many scroll-past impressions. |
| A partnership announcement | Proof you exist commercially, borrowed from whoever you partnered with. |
| A newsletter mention | Editorial selection. Somebody decided you were worth the space. |
| An advertisement | Reach at will. Weak alone, useful once the other encounters have happened. |
| A friend sending a link | The strongest of all, and the only one nobody can buy. More on that below. |
The variety is doing real work. Each source carries a different kind of authority, and a buyer who has met you through five of them has something closer to a rounded impression than a buyer who saw one thing five times.
Which sets a constraint on how the touchpoints get made. If all fifteen are the same paid message in different fonts, you have not built familiarity, you have run a frequency campaign with extra steps.
The encounters have to come from people who would say it anyway, in their own words, which is why creator selection and briefing matter more than the size of the buy.
What one launch actually delivers
Here is the arithmetic that reframes a launch, and it is worth doing before you plan one.
A launch that reaches two million impressions does not give two million people one touchpoint each. It gives a much smaller number of people several, and most of the audience one.
For a given individual in your actual buying market, a strong launch day is two or three encounters: they see one creator cover it, maybe a second, and perhaps the founder's own post.
Two or three, against a requirement of roughly fifteen.
It also explains a pattern every team notices and few explain: the signups that arrive three weeks after launch week, from people who saw it and did nothing at the time. Those are buyers who were at touchpoint two in August and reached fifteen in September through channels nobody was tracking.
Why this breaks your attribution
The multi-touch reality is the single biggest reason good distribution gets cancelled, and the mechanism is worth stating plainly because it happens in a meeting nobody thinks is about attribution.
Standard models assign a conversion to one event. Last-click gives full credit to the fifteenth touchpoint, which is almost always a branded search or a direct visit, and awards precisely nothing to the fourteen encounters that made the buyer type your name. First-touch inverts the error and credits whatever happened to be first.
So the channel that produced touchpoints two through fourteen reports close to zero, and the channel that caught the person at the end reports the whole conversion. A team acting rationally on that data cuts the first and scales the second, then watches the second stop working, because it was harvesting familiarity somebody else built.
Three practical defences.
- Record a baseline and measure the step, not the spike. The number that captures accumulated familiarity is where your floor settles afterwards. A launch that moves daily signups from 40 to 65 permanently has done something no per-campaign attribution will show you. The measurement scheme.
- Watch branded search above everything. It is the closest available proxy for how many people are somewhere between touchpoint one and fifteen, and it moves before revenue does.
- Ask, in free text. The single question about how somebody heard of you is the only instrument that catches a fifteen-touchpoint path, and the answers routinely name three separate sources in one sentence.
Why several drops beat one
If a buyer needs fifteen encounters and a launch delivers two or three, the strategy follows directly. You need multiple moments across months rather than one large one a year, and each should reach a partly overlapping audience so the same people accumulate.
This is the mechanism behind a campaign we can point to publicly. Higgsfield ran six releases through the network in four months, and every drop cleared the same floor, between 84 and 121 times a normal Higgsfield post for the four smaller releases.
Repeated moments in front of a partly overlapping audience is the fifteen-touchpoint model showing up in a launch calendar.
The corollary is uncomfortable and true. A company that launches once and goes quiet has bought a set of people at touchpoint two and then let them decay, and attention decays without contact. What compounds and what decays covers the shape of that.
The touchpoint nobody can sell you
Worth saying clearly, since we are in the business of selling the other ones.
The most valuable encounter on the list is a friend sending someone a link. It arrives with full trust, it is unmediated, and it converts at rates no purchased touchpoint approaches. It cannot be bought, manufactured or scheduled, and any agency implying otherwise is describing something else.
What you can do is make it more likely, which is a different and smaller claim. People forward things that are worth forwarding: an artefact somebody can inspect, a claim worth arguing about, a number that surprised you, something funny enough to be worth the social risk.
Designing for that raises the probability. Nothing guarantees it.
The honest framing of what amplified distribution buys you is this: it manufactures the encounters that can be manufactured, so that the ones that cannot have more chances to happen.
Planning for fifteen instead of one
Four steps, and the first one takes an afternoon.
- Count what you actually produce in a quarter. List every distinct kind of encounter a target buyer could have with you in ninety days. Most teams doing this find three or four, and the gap between that and fifteen is the plan.
- Fill the gaps by kind rather than by volume. If you have creators and advertisements and nothing else, the next thing to add is a podcast appearance or a partnership, not more of the two you have.
- Space them. Fifteen encounters in one week is a frequency campaign. The same fifteen across a quarter is familiarity, and the spacing is what lets memory form.
- Give the sources genuine reasons. Encounters from people who would have said it anyway compound. Encounters that read as bought get discounted the same way anything on a channel you control gets discounted.
The mental shift underneath all of it is small and hard: stop asking whether a campaign converted, and start asking how many people it moved one step closer. The second question is answerable, it is the honest description of what distribution does, and it is the one that stops teams cancelling channels that were working.
Questions we get asked
How many touchpoints does a B2B buyer need before signing up?
One of our clients measured it at roughly fifteen, and the important detail is that they have to differ in kind. Fifteen impressions of one advertisement is frequency and it produces fatigue. A founder clip, a creator mention, a podcast, a newsletter and a friend sending a link are five different encounters, and different sources are what turn repetition into familiarity.
How many touchpoints does a single launch produce?
Two or three for any given person, even when the launch reaches millions. Somebody sees two creators cover it on the same afternoon and that is most of what they get. This is the honest arithmetic behind why one launch a year underperforms and why several drops across a few months compound, since each one starts from a warmer position than the last.
Why does attribution make this channel look bad?
Because standard models assign a purchase to one event. Last-click credits the fifteenth touchpoint, usually a branded search, and awards nothing to the fourteen that made the buyer type your name.
First-touch does the reverse. Both are wrong in the same way, and teams cancel the channel that built the familiarity because the model cannot see it.