The Cold-Email Arms Race Is Starting to Work Against Sales Teams

Maverrik, a Plymouth-based social-selling training business, has launched what it calls a Human Selling Framework, arguing that B2B outbound is entering an unproductive contest: sellers use AI to produce more prospecting messages while recipients and email systems become better at ignoring, filtering or rejecting them.

The diagnosis is directionally credible, though the proposed framework and its claimed results are Maverrik’s own. The more useful question for commercial leaders is not whether cold email has suddenly stopped working. It is whether its economics are being distorted by a volume-first operating model that makes relevance, reputation and buyer attention progressively more expensive to secure.

Low reply rates are a measurement problem as well as a sales problem

The figures cited in the announcement should be treated carefully. Cold-email benchmarks vary considerably according to audience, sector, message type, data quality, sending infrastructure and—crucially—what counts as a reply.

Still, recent provider datasets point to a difficult environment. Sales.co reports that, across 1.29 million messages sent from October 2025 to August 2026, 0.70% produced a human reply and 0.24% an interested human reply. Woodpecker puts the median reply rate across its own 2025 platform data at 1.5%.

Neither dataset is a definitive picture of all B2B prospecting. Both are drawn from commercial platforms and should not be read as universal performance standards. But together they reinforce a practical point: a headline response rate can conceal automated replies, negative responses and activity with no plausible route to revenue.

That makes an indiscriminate push for more sends a poor remedy. It can raise the workload involved in list building, inbox management and follow-up without improving the quality of conversations reaching the sales team. It can also disconnect outbound reporting from the evidence that matters: relevant meetings, progression through the buying process, deal quality and retained revenue.

Deliverability has turned volume into a reputational issue

The technical constraints are becoming more important too. Google’s sender guidance requires organisations that send around 5,000 or more messages a day to personal Gmail accounts to meet authentication, unsubscribe and spam-rate requirements. It advises bulk senders to keep user-reported spam below 0.1% and avoid reaching 0.3%, warning that non-compliant traffic can face delivery disruption or rejection.

Those rules do not make legitimate B2B outreach impossible, and they do not apply neatly to every corporate inbox. Yet they illustrate a wider change in the commercial environment: sender reputation is part of the route to market. A sales process that depends on ever more domains, inboxes and message variants may be addressing symptoms of weak engagement while creating fresh operational and brand risks.

For communications leaders, this matters beyond the sales-development function. Prospects do not always distinguish between a badly targeted automated approach and the wider company behind it. An email that uses a superficial personal detail, misunderstands a business priority or arrives repeatedly from related addresses can undermine the credibility that marketing, customer success and senior executives are trying to establish elsewhere.

The stronger case for human input is better judgement

Maverrik’s response is to emphasise familiarity, understanding of buyer thinking and human-led personalisation, often using LinkedIn before direct contact. Its framework will be incorporated into workshops and programmes from 2027.

There is a sensible principle underneath the promotion. Research and familiarity should not mean collecting stray personal facts to make a template appear bespoke. They should mean forming a defensible view about whether a company has a relevant problem, whether the timing is plausible, who is affected, and whether the seller has a useful contribution to make.

That requires judgement. The salesperson needs to distinguish a real business signal—such as a strategic change, operational constraint or stated customer priority—from a generic account attribute. They also need the permission to decide that no message should be sent when the case is weak.

LinkedIn’s own sales guidance has long framed trust-building as work that begins before a connection request or initial call, through research, a credible professional profile and relevant contribution to the buyer’s world. That is not proof that social selling will generate the double-digit response rates Maverrik says it has observed; the company has not published enough methodology for that claim to be assessed. It does, however, support the broader idea that a buyer’s prior impression shapes whether direct outreach receives attention.

Use AI behind the conversation, not in place of it

The useful division of labour is less dramatic than an anti-AI argument. AI can help salespeople summarise public information, prepare account briefs, update CRM records, surface patterns in customer feedback and reduce administrative effort. These are tasks where speed and consistency can free time for a person to investigate a genuine commercial hypothesis.

The final outreach decision should remain more demanding. Teams should be able to explain why a prospect was selected, which buyer need or organisational change makes the contact relevant, and what evidence would show that the approach is improving outcomes rather than merely sending more messages.

That also creates a better test for outbound investment. Instead of asking how cheaply a team can create 10,000 personalised-looking emails, leaders can ask how reliably it creates a smaller number of informed, welcome and commercially meaningful conversations.

Maverrik’s announcement is therefore best read not as an obituary for cold email, but as a warning about false efficiency. Automation can scale activity rapidly. It cannot, by itself, establish relevance or earn trust.

---

This article is based on information distributed through Pressat. It has been edited by Expert View Media for clarity, context and length.