Rewrite your offer before scaling outbound

Why the offer comes before volume

Volume multiplies whatever you send. If the offer fits the buyer, more volume gives you more meetings. If it does not, you get more silence, more complaints and a worse sender reputation. Woodpecker's data on 7,676 mailboxes shows the cost of pushing volume: reply rates fell from 2.0 percent at 1 to 20 emails a day per mailbox to 0.5 percent at 101 to 200, and to 0.2 percent at 1,001 to 2,000 (Woodpecker; third-party data, checked 18 October 2026). The same page shows campaign size working the same way: independent companies had a 2.7 percent median reply rate at 0 to 50 prospects per campaign and 0.5 percent at 10,001 or more. Those are correlations from one vendor's customers, and they point the same way as the narrow, segmented campaigns we build.

Relevance is the other pressure. Gartner's survey of 632 B2B buyers (August to September 2024) found that 73 percent actively avoid suppliers that send irrelevant outreach (Gartner; checked 18 October 2026). Gong's article on prospecting email data reports that 87 percent of buyers say the sales emails they receive do not address a relevant challenge in their organization (Gong Labs; checked 18 October 2026). If you cannot state the offer in two sentences, volume will not fix it.

A rewrite is cheap compared with a burnt domain. The ordering we use is: audience first, offer second, copy third, volume last. The ICP step is covered in building an ICP for outbound, and this playbook picks up at the offer.

Score the offer before you send it

An offer is what the buyer gets for replying, plus the cost to them of finding out whether it is useful. April Dunford's definition of positioning, "how our product is different and better than alternatives for a particular set of customers", is a useful test of the underlying claim (aprildunford.com; checked 18 October 2026). The scorecard below checks whether that claim has been turned into something a stranger will accept. It is our own working tool, not a published standard.

Score each criterion 0 (absent), 1 (partly there) or 2 (clear). Maximum 14.

Scores of 11 or above are ready for a small test. Between 8 and 10, rewrite the lowest-scoring rows first. Below 8, the offer is probably a description of your service, and no amount of sending will improve it.

Here is an illustration of one offer rewritten (example wording only; claim only work you have done). The first version is a typical agency pitch: "We help B2B brands grow with data-driven marketing, and we would love 20 minutes to learn about your goals." It scores about 3: no named buyer beyond "B2B brands", no deliverable, a 20-minute ask. The rewrite: "We looked at the pricing page and checkout flow of three e-commerce brands in your category and found two fixes that each take under a day. I can send the one-page version for yours. Want it?" It names a buyer type, offers a deliverable, costs one word and delivers in days. It would score 10 or 11, with proof still to add.

How to test an offer without fooling yourself

Test in two stages: read first, count second.

Stage one is qualitative. Choose one segment and send each of two or three offer variants to a few hundred contacts from the same list, from the same inboxes, on the same days. Do not wait for statistical significance. Read every reply and sort it by reason: interested, not now, wrong person, already have a solution, not relevant, unsubscribe. A pattern in the reasons tells you more than the reply rate. If most "wrong person" replies come from one title, the audience is off. If the interested replies mention the same line of the email, that line is your offer.

Stage two is quantitative, and it needs more volume than most teams expect. The table below uses the standard normal approximation for comparing two proportions at 5 percent significance and 80 percent power. It is our arithmetic, not a published benchmark, and calculators such as Evan Miller's take the same four inputs (baseline, minimum detectable effect, significance and power) and may return slightly different numbers.

Two lessons follow. A large difference in offer quality shows up with a few hundred sends, but a small one does not. And positive replies, the metric you care about, need several times more volume to test than total replies. When you cannot reach those sample sizes, use the reply-reason sort from stage one and make decisions on large, obvious gaps.

Fix the sample size before you start. Evan Miller's guidance on A/B testing is that repeatedly checking results and stopping when they look significant makes the reported significance invalid, so decide the sample in advance and wait until the test is over (Evan Miller; checked 18 October 2026). The metrics worth tracking at each stage are in outbound metrics and benchmarks.

Case: an offer rebuilt around a new ICP

The e-commerce marketing agency in #gtm-12 came to us with an offer that no longer matched its market. Earlier appointment-setting agencies had produced little. We rebuilt the offer around a newly identified ICP, coached the founder on how to present it, and launched tailored cold email and LinkedIn campaigns with lead handling ready for the volume. The result was 40+ qualified appointments and five clients closed in a single week (case #gtm-12).

What changed was the offer and who it was for. Those are the parts the scorecard measures: named buyer, deliverable and clarity.

A second case shows the same logic when the problem is stalled deals rather than empty calendars. A GTM engineering agency in #gtm-9 had stalled deals, unclear positioning and client churn risk. We reworked the strategy, tightened messaging and the proposal flow, supported closing calls and rebuilt trust with existing clients. More than 10 stuck deals closed and 12 or more existing clients stayed, which steadied revenue (case #gtm-9). Its engagement type was deal recovery and retention, and it belongs here because the repair work was positioning, messaging and the proposal flow.

Two other cases show an offer that is easy to accept. For Fema Tape, a sample sat in every message: 850 firms were contacted in month one, with 32 positive replies against a target of 15 (Fema Tape case study). Fema Tape sells a physical product, which makes a sample natural. A service has to find its own version.

A rewrite process you can run in a week

1. Pick the segment. One title at one company type. If you cannot, go back to the ICP.
2. Write three offers for that segment, each with a different deliverable (for example a teardown, a sample, a short list of accounts).
3. Score each on the table above. Rewrite anything under 8.
4. Send each of the top two or three to 200 to 400 contacts. Keep the list, inboxes, days and send times the same.
5. Wait about a week, since 93.9 percent of replies arrive within 72 hours in Smartlead's dataset of 850 million emails (Smartlead; checked 18 October 2026), then read every reply and sort by reason.
6. Keep the offer with the best positive replies from the right buyers. Rewrite the lowest-scoring criterion and test the next version against it.
7. Set a gate for scaling: you can name the buyer, positive replies are coming from that buyer, and you can handle the calls you will book.

The last gate is capacity. In #gtm-8, a PR and brand agency pitched PR and design as one service, with separate messages for startups and multinationals. It booked 130+ meetings, 35 of them in the first month, and hired two people to handle the volume (case #gtm-8). Know how many calls you can take before you scale.

If you want help with the offer, positioning or the first test, our GTM strategy service starts there. The cold email copy that carries the offer is covered in how to write cold emails that book meetings.

Questions

How do I know if my offer is weak?

Read the replies from a few hundred sends. If most say wrong person, not relevant or already have a solution, the audience or the offer is off. If positive replies mention the same line, that line is your offer.

How many emails do I need to test an offer?

For large differences, 200 to 400 contacts per variant gives a readable signal. To detect a move from a 2 percent to a 4 percent reply rate with standard statistical settings, you need about 1,140 emails per variant. Positive-reply tests need several times more.

Should I change the offer or the list first?

Audience first. An offer cannot rescue the wrong buyer. Once the list is right, test the offer, then the copy, then raise volume.

What makes a good cold email offer?

A specific deliverable (a sample, a teardown, a short list), a named buyer, and a reply that costs one word. Asking for a 20-minute call from a stranger is the weakest form.