Most demand programs for deep-tech companies are sized for a market that never existed, and the fix starts with counting the real account list.
Most deep-tech and robotics sellers have a few hundred real buyers. Count the named accounts that fit your profile and that sales would accept, then the buying-group contacts in each, and let that number decide whether to run broad paid demand, account-by-account coverage, or a coverage-building phase first.
A robotics company selling warehouse automation has a few hundred potential buyers, while most demand-generation programs are sized for tens of thousands. That arithmetic, checked against the real account list rather than assumed from a category benchmark, is why so many programs built for AI, robotics and deep-tech companies fail before the first campaign finishes learning: the program was sized for a market that never existed.
Most paid platforms, most ABM tools and most demand-gen frameworks were built for markets like project management or marketing software, categories with tens of thousands of qualified buyers and enough volume for a platform's algorithm to run a real learning phase, test creative variants and converge on a working audience. A warehouse automation vendor, a physical AI company or a sensing hardware maker selling into named industrial accounts doesn't have that volume. The buying group per account is small, the number of accounts that can plausibly buy is small, and a platform told to find a pattern in a four-hundred-account universe behaves the way any statistical system behaves with too little data: it never exits the learning phase, it spends the budget searching for a pattern the sample size can't support, and the reported results look inconclusive because they are.
LinkedIn's own rules make the arithmetic concrete. An ad set needs at least 300 matched members before it can run, LinkedIn recommends 50,000, or 300,000 for Sponsored Content, and the count is of members it could match, so an uploaded list shrinks after matching. Four hundred accounts with three or four reachable people in each produce a list of roughly 1,200 to 1,600 people before matching, close to the floor and far below the recommendation, which is why the learning phase never ends.
You can see the gap in your own numbers. Run the firmographic count for your product category against the account list your sales team would actually accept, then compare it with the audience size your ad platform's targeting tool reports for the same category. The difference between those two figures is where wasted budget lives.
Paid media can still work in a small market, as long as the program is built for the market's real size. A four-hundred-account universe with a defined buying group per account can carry a program, just a different one from what a forty-thousand-account market would run. Coverage of the whole buying group across those four hundred accounts, mapped by name, becomes the unit of work in place of reach across an anonymous audience, and that changes the structure of the campaign, the budget pacing and what counts as a result worth reporting.
For a revenue leader briefing a program, this means starting with the count before the creative. Before writing an ad, a sequence or a content calendar, get the real number: how many named accounts actually fit the ideal customer profile, how many buying-group contacts sit inside each one, and how many of those contacts are already reachable versus still unidentified. That number decides whether the right motion is broad paid demand, tight account-based coverage, or a coverage-building phase before either makes sense, and it decides how a platform's learning-phase budget should be spent, if it should be spent there at all.
The size of your addressable market doesn't improve with better copy or a bigger budget, so you check it before deciding either. The same logic applies past robotics to any company selling a system integrated into a small number of large operations, from industrial software to specialised sensing to applied AI sold into a narrow vertical. Pull the actual named-account list for your category this week and count it, because that single number changes what every campaign built on top of it should look like.
Each figure links to the page it comes from, so you can check it before you repeat it.
| What the evidence says | Source |
|---|---|
| LinkedIn requires at least 300 member accounts for an ad set and recommends at least 50,000, or 300,000 for Sponsored Content. | LinkedIn Marketing Solutions Help |
| For Sponsored Content and Sponsored Messaging, LinkedIn suggests a minimum audience of 300,000. | LinkedIn Marketing Solutions Help |
| On average 13 people inside an organisation are involved in a B2B buying decision, and 89% of purchases involve two or more departments. | Forrester (4 December 2024) |
| LinkedIn's B2B Institute estimates that 95% of potential B2B buyers are not ready to buy at any given moment. | LinkedIn B2B Institute |
Each answer stands on its own, so you can forward one to a colleague, and an assistant asked the same question can quote it as it is.
There is no universal number. Count the named accounts that fit your profile and that sales would accept, then match the program to that count, which for many deep-tech sellers is a few hundred and suits account-by-account coverage.
LinkedIn requires at least 300 matched members for an ad set and recommends 50,000 or more, or 300,000 for Sponsored Content. The count is of matched members, so an uploaded list usually shrinks after matching.
Platform algorithms need volume to learn which audience responds. With a few hundred accounts they rarely leave the learning phase, so the budget goes into searching for a pattern instead of reaching the buying group.
Run the firmographic count for your category, filter it to the accounts sales would actually accept, then count the buying-group contacts per account and how many of them are reachable. Compare that with the audience size your ad platform reports to see where budget is being wasted.
Check whether LinkedIn ads can serve your named account list, what a month of reach costs at your CPM, and what to report to your founder in place of a lead count.
Tell us what your team sells, who buys it and where deals stop moving. You'll get a written reply saying which way of working fits you, or whether a Diagnostic makes sense, and if neither does, the reason.