The first spam

On 3 May 1978, Gary Thuerk, a marketer at Digital Equipment Corporation, wanted to reach computer users on America’s west coast. He had his project manager copy roughly 400 addresses by hand from the printed directory of ARPANET, the government network that came before the internet, and sent them all one message inviting them to product demonstrations in California.

According to David Kalat’s history of the episode, the demonstrations went on to be credited with over $12 million in sales. They also earned DEC a complaint from the agency running the network, which called it a “flagrant violation” of a network meant for official government business. Thuerk promised never to do it again.

It’s usually told as a funny footnote: the first spam. But it holds the whole tension of email marketing in one episode. The message reached people cheaply and it worked. And it annoyed enough of them that the network moved to stop it.

Which means the question for anyone running nurture today isn’t whether email works. It’s how much is too much - and whether your dashboard can tell you.

Why this matters for leaders

Every nurture programme has an argument about cadence. Sales wants more touches on their accounts. Marketing operations worries about unsubscribes and deliverability. Someone quotes a “rule of seven”, which I couldn’t trace to any source. And the decision usually gets made by opinion, then measured on open rates.

The research on email is mostly from retail and consumer businesses, and I’ll come back to what that means for B2B. But some of it is experimental, and it’s clearer than the argument usually is.

So I’ve broken it into three parts: why there’s a right number of emails, why the dashboard is misleading you, and what actually moves people.

1. There is a right number

Xi Zhang, V. Kumar and Koray Cosguner modelled email and purchase behaviour for a US home improvement retailer. The number of emails had a non-linear effect on both short and long-term profit. In their analysis, the best number was seven. Sending four instead would have lost 32% of lifetime profit per customer - and sending ten would have lost 16%.

Both directions cost money. Too few and you leave value on the table; too many and you wear people out. An earlier study by Kumar, Zhang and Anita Man Luo found that marketing intensity affects how long people stay opted in, and that firms could extend that time by allocating contacts more strategically.

My reading is that there’s no universal right number. Seven was one retailer’s answer. But there is a right number for your audience, and it sits somewhere between the two ways of getting it wrong - which means you can only find it by testing more than two levels.

2. The dashboard is lying

The same retailer study found something more uncomfortable: customers who were active on email weren’t necessarily active buyers, and vice versa. The people opening your emails and the people buying from you overlap less than the dashboard implies.

And opens have got less reliable since. Apple’s Mail Privacy Protection, introduced with iOS 15 in 2021, prevents senders from seeing whether a message was opened. A nurture programme managed on open rates is now partly managed on noise.

Email’s effect also fades fast. Daniel Zantedeschi, Elea McDonnell Feit and Eric Bradlow studied a retailer that randomly held out customers from every campaign, which let them measure what each channel actually did. Catalogues had a much longer-lasting effect on purchases than emails. And targeting the customers most responsive to advertising increased predicted returns by about 70% compared with targeting on past purchase behaviour.

For me, that combination is the problem with most nurture reporting. It counts the wrong behaviour, through a measure that’s now partly hidden, for a channel whose effect is short. I wrote in spread it thin about why spacing matters for exactly this reason.

3. What moves people

The most useful email experiment I know of is also the simplest. Navdeep Sahni, Christian Wheeler and Pradeep Chintagunta ran randomised experiments with three companies, sending millions of emails. In the main one, adding the recipient’s name to the subject line raised opens from 9.05% to 10.80%, raised sales leads by 31%, from 0.39% to 0.51%, and cut unsubscribes by 17%.

A name tells you nothing about the product. The researchers’ explanation is that it made people put more effort into reading the rest of the message. That’s a useful reminder that “personalisation” doesn’t have to mean dynamic content or a recommendation engine. Feeling like the email is meant for you does a lot of the work.

There’s a cost on the other side too. Koen Pauwels and Allen Weiss found that an online content firm’s conversion email blasts raised subscription revenue at the expense of advertising revenue. Every email has a cost that doesn’t show up in the campaign report.

What good looks like

A few checks I’d use on any nurture programme:

• Cadence tested, not debated. At least three levels, with a quarter to run.
• A permanent holdout. A random slice of leads that gets no nurture, so you know what it adds.
• Meetings and pipeline as the measure. Opens as a diagnostic at most.
• Unsubscribes counted as a cost. In the test results, not a footnote.
• Something personal in every email. A name, a sender who’s a person, a reason it’s for them.
• No “rule of seven” slides. Find your own number.

For me, yes and no

There’s a fair challenge here. Isn’t this all retail email? A B2B buyer reading a nurture sequence isn’t a homeowner looking at a paint promotion.

I agree, and it’s the biggest limit on this essay. The profit model and the holdout study are from retailers, and the personalisation experiments were with three companies selling a range of products, measured partly on sales leads. I couldn’t find peer-reviewed field experiments on B2B nurture cadence.

But the mechanisms seem general to me. Both too few and too many emails cost something. Email activity isn’t buying activity. Opens can’t be trusted as they once could. And small human touches move response. For me, the B2B version of the lesson is the same as the retail one: stop arguing about cadence and test it, against the outcome you actually care about.

Where to start on Monday

1. Change the headline metric. Report your nurture programme on replies, meetings and pipeline, and move opens to a diagnostic tab.
2. Count the cost. Add unsubscribes and spam complaints per thousand contacts to the same report.
3. Check who’s opening. Compare your most engaged email contacts with the accounts that created pipeline last quarter, and note the overlap.

Three tests to run

1. There might be a sweet spot below your current cadence, or above it. Randomly split a nurture audience in one country into three cadences - say fortnightly, weekly and twice weekly - for a quarter. Measure meetings, pipeline and unsubscribes per thousand contacts.
2. A permanent holdout could show nurture is worth less, or more, than attribution says. Hold back a random 10% of new leads from all nurture for two quarters. Compare the rate of sales-qualified leads and pipeline per lead with the nurtured group.
3. Emails from a named seller might outperform branded ones. Randomly send the same content from the account owner’s name or from the brand. Measure reply rate, meetings and unsubscribes.