Almost every failed outbound programme fails in the same place. The first email goes out, a handful of people reply, and the sender concludes the list was bad or the copy was weak. Neither is usually true. What happened is that the campaign ended at touch one, in an inbox where the recipient was in a meeting, on holiday, or looking at something else that morning.
Why five, and why the gaps are not equal
A single send is mostly luck. A properly constructed sequence reliably produces several times the reply volume of its first email alone, and the extra touches cost nothing but the discipline to write them. Past five or six the curve flattens hard while the complaint rate does not, which is the whole argument for stopping there.
The spacing that works widens as it goes: day zero, plus two, plus three, plus four, plus five. Eleven business days from first contact to break-up. Early follow-ups sit close together because they are fighting inbox timing, not opinion. Later ones need distance so they read as persistence rather than desperation — a difference of days, not words.
Three non-negotiable rules. Business days only, because prospecting on a Saturday reads as automation regardless of what you wrote. Send inside the recipient’s working hours rather than yours — country and lang sit in every store record, so a three-in-the-morning delivery is a choice, not an accident. And stop on any reply at all, including negative ones: nothing destroys goodwill faster than touch three landing after somebody answered touch two.
Touch one: one observation, one implication, one question
The opening email has three jobs. Prove you looked. Turn what you saw into a consequence the merchant cares about. Ask something easier to answer than to ignore. Fifty to a hundred words, and the test is mechanical: send it to your own phone, and if you have to scroll, cut. What it must never do is request a meeting — touch one exists to pre-qualify and create curiosity, not to sell.
Delete the throat-clearing. “I hope this email finds you well.” “I am writing to you because.” “I know you are busy.” “My name is X and I am Y at Z.” All of it is either in your signature or is filler announcing you had nothing to open with. Then count the word “I” against the word “you”. If the first wins you have written a brochure, and the fix is usually one word: “how we could get you those results” becomes “how you could get those results yourself”.
The observation has to be something the merchant did not tell you. Store-level facts that work as openers: the platform they run, what is and is not in the detected app stack, catalog size and the direction it has moved, the shape of the traffic curve across the monthly history, the Google rating and review count, whether the profile is claimed, the payment and carrier configuration, the languages the storefront runs in. A merge tag is not personalisation. A fact they did not publish in a press release is.
Keep the ask short. Seven words is the target and fifteen is the ceiling, because a call to action that wraps onto a second line does not get read. Two forms outperform the standard meeting request at the cold stage: the interest probe — may I send you what I found? — and the conversation opener — how are you handling this at the moment? Both are answerable in one line, neither costs the merchant a calendar slot, and both land you in a thread rather than a diary.
Touch two: the bump, two days later
The shortest email in the sequence and the easiest to get wrong. Twenty-five to forty words, sent as a reply in the same thread so the original sits directly underneath.
Never write “did you see my email?” It contains no new information and it quietly makes the recipient responsible for your outcome, which is a strange thing to do to a stranger. Instead, add one genuinely new data point and re-ask the same question in fewer words than before.
The shape is: one more thing I noticed, here it is, same question. Three seconds to read, it moves the thread back to the top of the inbox, and it proves you have more than one true thing to say about their business. That last part is what separates a bump from a nag.
Touch three: the artifact, three days later
This is the touch that separates operators from senders, and it is the one most people skip because it costs real time. New thread, new subject line, and it contains something you actually built for this specific merchant.
What an artifact looks like in this market: a one-page audit of their Google Business profile with the wrong category circled, an annotated teardown of everything a new subscriber receives in their first four days, three creative angles written against their visible best-seller, a keyword gap list against two named competitors. Twenty minutes of work, immediately usable whether or not they ever hire you.
Position it as already done, with nothing attached. “Made this for you. No catch and no call needed — if it is useful, use it.” The instant there is a condition, it stops being a gift and becomes a lead magnet, and merchants can smell that difference from the subject line. Sequences that give away real work consistently outperform sequences that describe work they would be willing to do.
Touch four: change the angle, not the volume
If three touches on one problem produced silence, the most likely explanation is that you picked the wrong problem. Touch four exists to test a second hypothesis, not to repeat the first one more loudly.
Store-level data makes this possible because it hands you more than one true observation per merchant. You opened on the flat traffic curve; touch four opens on the unclaimed Google profile, the missing email capture, or the single shipping carrier under a growing catalog. Same sender, different door.
This is also the right place for proof, and proof means mechanics rather than testimonials. A sentence describing how a mechanism behaves — what usually happens to abandoned-cart revenue when a flow is switched on, what usually happens to map visibility when a primary category is corrected — is a claim you can defend. Never invent a customer, a logo, a percentage or a case study. The one thing outbound cannot survive is being caught making things up, and merchants talk to each other.
Touch five: the break-up that does not guilt-trip
The final touch converts far better than most people expect, because it is the one email in the sequence that removes pressure instead of adding it.
The version that works hands the recipient an easy classification. Closing the loop on this — is it (a) not interesting, (b) not the right person, or (c) not the right time? A single letter is a perfect answer. Merchants reply because it costs one keystroke, and every possible answer is useful to you: (b) produces an internal forward, (c) produces a date to come back on, and (a) produces a clean removal that protects your sending reputation.
What to avoid: “I guess you are not interested”, which is passive aggression in a business suit, and “this is my final attempt before I remove you from my list”, which is a threat nobody has ever feared. A light, self-aware closing line is fine if the segment tolerates it, so check who you are writing to first. And one structural trap catches almost everybody: the break-up still needs an ask. A farewell email that hands over something valuable and then signs off without a question feels generous and produces nothing. Reciprocity does not generate replies on its own. Questions do.
And use the PS. Everybody reads it, which makes it the most valuable line in the email and the most commonly wasted. Do not put a second call to action there. Put something with no ask attached: a useful link, one honest joke, or a referral request — if this is not for you but you know somebody running a store in this shape, I would happily take the name.
Subject lines, volume, and the boring infrastructure underneath
Subject lines should be short, lowercase and honest. Three words beats five, and one word often beats three — question, video, audit, strategy — because a bare word reads like something a colleague or a supplier sent, while anything carrying a merge tag now announces itself as prospecting on sight. Match the word to what is actually inside: if the email contains a video, the subject is the word video. Stay out of the marketing-newsletter register entirely, because nobody subscribed to you.
Draw the distinction on the “Re:” prefix carefully. Faking one on a first contact to invent a shared history is deception and it poisons the thread before anybody reads a word. Leaving the subject field empty on a follow-up so the tool threads it under the original is not — that is correct threading, and it is what keeps touch two attached to its context. Trickery buys opens and loses deals.
Volume needs a ramp. A fresh sending domain should start around twenty emails a day and climb by roughly ten a day each week towards a ceiling of forty to fifty per mailbox. Follow-ups count against that ceiling: a five-touch sequence at ten new prospects a day settles near fifty sends a day once the pipeline fills, so plan your mailbox count backwards from your list size instead of discovering the limit by being throttled.
Then the unglamorous parts: authenticate and warm the domain for at least two weeks before you use it, route every reply to a human within the working day, and make the unsubscribe real and instant. None of this wins deals. All of it decides whether the deals you would otherwise have won ever reach an inbox.
Measure per touch, not per campaign
The most common analytical mistake is judging the sequence as a single unit. Attribute every reply to the specific touch that produced it, because a sequence averaging poorly is usually three good emails carrying two bad ones. In a well-built five-touch programme, expect touch one to produce the largest single share, touches two and three together to produce roughly as much again, touch four to be the weakest, and touch five to surprise you. If touch five produces nothing, it is written as a guilt trip. If touch three produces nothing, your artifact is not actually useful to the person receiving it, which is a harder truth and a more valuable one.
Change one touch at a time, and test in small batches. Four parallel mini-campaigns of a hundred addresses teach you more in a week than one two-thousand-address blast, and being wrong costs far less sending reputation. Separate reply rate from positive reply rate while you are at it, and watch the calendar as well: a real share of prospects book a slot without ever answering an email, so judging on replies alone kills sequences that were quietly working.
On targets rather than promises: treat anything above five percent reply on a cold, tightly targeted list as working, and a double-digit rate as evidence that your targeting and your artifact are both right. Aim at those, do not expect them in week one, and distrust anybody quoting guaranteed rates. Silence is not permanent either — a store that ignored five touches in March is a legitimate target again in the autumn with a new subject and a different angle.
Finally: the leverage is not in the sequence. Five touches against four hundred well-chosen stores beats five touches against four thousand random ones, and takes less of your week. Spend the saved time on the list and the artifact — the two things nobody can copy out of your email.