Most advice about sales follow-up is advice about stamina. Send more touches. Do not give up after one. "80% of sales are made on the fifth to twelfth contact." The message is that deals are lost by sellers who stop too soon, and the fix is persistence.

There are two problems with that framing. The first is that the most quoted persistence statistics in sales appear to be invented. The second is more useful: when you look at the evidence that does exist, the real failure in follow-up is not a shortage of messages. It is a shortage of messages that carry anything. Sellers under-ask when they should ask, over-send when they have nothing to say, and rarely leave the buyer holding something that moves the decision forward inside their own organization.

This article examines what the evidence supports about follow-up, where it stops, and offers a simple working test for whether the next message deserves to be sent.

The famous follow-up figures usually run like this: 48% of salespeople never follow up, only 10% make more than three contacts, 2% of sales close on the first contact, and 80% close on the fifth to twelfth. They are typically credited to the "National Sales Executive Association."

In 2014, two writers independently tried to trace the source. Adam Honig, writing in CustomerThink, concluded: "These purported statistics are completely made up! There is no National Sales Executive Association, nor did they publish any study" [1]. Two months later, Stewart Rogers at VentureBeat searched business registries, the Better Business Bureau network and IRS tools and reached the same verdict: "These stats are completely made up, and the National Sales Executive Association doesn't exist" [2]. Rogers noted that of the more than 300 references he had seen online, only two explained that the numbers were bogus.

This matters beyond pedantry. A fabricated statistic that says "almost every sale needs five to twelve touches" encourages a specific behavior: volume, applied evenly, regardless of what each touch contains. If the premise is false, the behavior it justifies deserves a second look.

Better, though still imperfect, evidence comes from vendor data on real outreach. Belkins, a B2B lead generation agency, analyzed 7,530,489 cold emails sent through its managed campaigns in 2025, broken down by sequence step and meeting outcome [3]. This is vendor data from one firm's client programs, not a controlled study, but it is large and recent.

Three findings stand out.

First, each additional email earns less. The first email in a sequence had the highest per-step reply rate (0.59%), falling to 0.30% by step 3 and 0.28% by step 6. Belkins describes step 6 as showing "minimal incremental value" and recommends three to five steps for most B2B sequences.

Second, and against the "one and done" instinct, later steps still matter in aggregate. Steps 2 through 6 together produced 58.6% of all replies, and steps 3 to 5 drove 53.5% of email-sourced meetings. The appointment rate stabilized at 0.2% from step 3 onward, twice the rate of steps 1 and 2.

Third, context changes everything. On LinkedIn, messages to people the sender was already connected with earned a 12.22% reply rate, against an overall email reply rate of 0.45% [3].

Read carefully, this is neither a case for relentless persistence nor a case for giving up. It is a case for a short, deliberate sequence where each step has a job. It also carries an important limit: the prospects still in a sequence by step 3 are, by definition, the ones who have not opted out. Higher conversion at later steps partly reflects who is left, not the magic of the third email. And cold sequences are only one kind of follow-up. Most of the follow-up that decides B2B deals happens after a first meeting, between stakeholders, in the weeks when a deal is quietly alive or quietly dying.

If volume is not the core problem, what is? One strand of psychological research suggests sellers systematically misjudge how their follow-up will be received, and the error runs in a consistent direction.

In a series of six studies published in the Journal of Personality and Social Psychology, Francis Flynn and Vanessa Lake (now Bohns) tested whether people underestimate how often others will comply with direct requests [4]. In the first three studies, participants "underestimated by as much as 50% the likelihood that others would agree to a direct request." In a field version with charity fundraisers, volunteers predicted they would need to approach 50% more people than they actually did [5]. The mechanism matters for sellers: people asking for something focused on the cost to the other person of saying yes, and underweighted the social cost to the other person of saying no.

A second line of work points the same way. Peggy Liu, SoYon Rim, Lauren Min and Kate Min ran 13 preregistered experiments with nearly 6,000 participants, published in 2022 in the Journal of Personality and Social Psychology: Interpersonal Relations and Group Processes [6]. People who initiated contact consistently underestimated how much the recipient appreciated it. The gap was larger when the outreach was more surprising and when the ties were weaker [6][7].

These findings help explain a familiar pattern. A seller has a good first meeting, then hesitates to follow up because they do not want to seem pushy, and the deal drifts. The research suggests that hesitation is often miscalibrated.

The limits are just as important. Flynn and Lake studied mostly face-to-face requests, where saying no carries a real social cost. In an inbox, the cost of ignoring a message is close to zero, so the protection that made people comply in person largely disappears. Liu and colleagues studied contact between people with a positive history; Liu herself cautioned that the results may not hold after a falling-out [7]. Neither study tested a sales sequence to a stranger. What the research supports is narrower and still useful: when you have a legitimate reason to reach someone you have already had a good interaction with, you are probably underestimating how welcome a direct, specific message will be. It does not license the sixth "just bumping this to the top of your inbox."

The most undervalued follow-up in B2B is the one sent within a day of a meeting: the recap. It is often treated as courtesy. The evidence suggests it does real work.

Memory for new material decays fastest in the first hours after learning. Jaap Murre and Joeri Dros replicated Ebbinghaus's classic forgetting curve in 2015, with one subject spending about 70 hours learning and relearning lists at intervals from 20 minutes to 31 days; the results were "similar to Ebbinghaus' original data" [8]. The limits are obvious: one participant, nonsense syllables, and a meeting about a business problem is far more meaningful than a list of syllables. The study does not tell you how much of your discovery call a buyer will remember. It supports a more modest point: recall drops quickly and unevenly, so the version of the meeting that each attendee carries away will begin to diverge within a day.

In a buying group, divergence is the danger. The people who were in the room remember different priorities, and the people who were not in the room hear a retelling from someone who never had to explain your offer before. A recap that states what was heard, what was agreed and what happens next gives everyone the same record and gives the buyer a document they can forward without rewriting.

There is also evidence that specificity about next steps changes follow-through. Peter Gollwitzer and Paschal Sheeran's meta-analysis of 94 independent tests found that implementation intentions, plans that spell out the when, where and how of an action in an "if situation Y, then I will do X" form, had a medium-to-large effect on goal attainment (d = .65) [9]. These studies concern individuals pursuing their own goals, not buyers keeping commitments to a vendor, so the transfer is an inference rather than a finding. But the logic carries: "let's reconnect soon" is a goal intention; "Priya will share the security questionnaire by Thursday and we will review it on the 14th at 10:00" is an implementation intention. Only the second gives the follow-up something to check against.

Gartner's research on buyer enablement offers the buyer-side view. In a 2018 release, Gartner reported that "customers who receive helpful information that eases the purchase process are three times as likely to buy the bigger, more expensive option, with less regret" [10]. Its customer confidence research describes information that helps customers advance as producing a 2.8x change in the likelihood of a purchase ease experience and a 3x change in the likelihood of a high-value, low-regret deal, and reports that 77% of B2B buyers described their latest purchase as very complex or difficult [11]. Gartner does not publish the sample size or survey year on these pages, so treat the multiples as directional analyst findings rather than precise estimates.

The implication for follow-up is direct. A buyer between meetings is not waiting to be reminded you exist. They are trying to complete buying jobs: explaining the problem to a colleague, comparing options, getting security or finance comfortable, building a case. A follow-up that helps with one of those jobs is welcome in a way that a check-in is not. Gartner's distinction between prescriptive advice (what to do and how) and practical support (help completing discrete tasks) is a useful filter for what to send [10].

Pulling this together, the useful question about any follow-up is not "how many touches is this?" but "what does this message carry?" A follow-up passes the Carry Test if it carries at least one of four things the buyer can use. If it carries none, do not send it yet.

1. A Record. A shared, accurate account of what was said and agreed: the problem in the buyer's words, the priorities, the open questions, the owners and dates. The recap is the obvious example. Its test is whether the buyer could forward it to an absent colleague without editing.

2. A Resource. Something that helps the buyer complete a buying job they are actually facing: an answer to the question they raised, a peer reference matched to their situation, a short comparison of options, a draft of the business case section they need to write. The test is whether it saves the buyer time on a task they have, rather than adding reading they did not ask for.

3. A Request. A specific, direct ask tied to a concrete next step: a decision, a meeting with a named person, a document by a date. The Flynn and Lake research suggests you are likely underestimating the chance of a yes, so ask clearly. The implementation intention research suggests the ask works better when it names when, where and who.

4. A Release. An honest, low-pressure exit that respects the buyer's time: "If this has moved down the list, tell me and I will come back next quarter." Belkins' own recommended final step is a low-commitment ask with a long horizon [3]. A good release often produces a clarifying reply, and when it does not, it ends the sequence with your credibility intact.

Messages that carry none of these, the "checking in," "circling back" and "any update?" notes, are the ones that train buyers to ignore you. The Carry Test does not cap the number of follow-ups. A complex deal may justify a dozen messages over two months if each one carries something. A cold prospect may justify three or four.

Follow-up is not a test of the seller's stamina. It is a test of whether each message makes the buyer's decision easier. Persistence is a virtue only when it has cargo.