You have your ICP written down on a page and the sales team has signed it off. Someone suggests hiring another SDR and raising daily sends. It looks like the logical next step, but nothing in that document has been tested yet: it is a hypothesis about who will buy from you. If you scale it as it is, volume will multiply what works and what doesn’t, and you won’t be able to tell one from the other. Outbound itself is a cheap way to test that hypothesis. Here is a method to validate your ICP with campaign data: how to set up tests by segment, how much to send so the result means something, which metrics to watch, and when to scale, adjust or drop each segment.
Starting point: an ICP written as a hypothesis
We start from an ICP that is already defined. If you don’t have one yet, begin with how to define the ideal customer profile and its role in lead generation with an ideal customer profile.
What changes here is the attitude with which you look at that document. Every ICP criterion (industry, size, region, business model, timing signals, exclusions) is a bet. Validating means putting each bet in front of real prospects and letting their answers tell you whether it holds. In SaaS there is also a nuance between the customer who fits on paper and the one who actually gets value from the product; we explain it in this comparison of ECP vs ICP in SaaS and the differences between each profile.
Why validate before scaling
Scaling an untested ICP has three costs that show up late.
The first is team hours. An SDR who writes to accounts that don’t fit researches useless companies and drafts messages nobody answers. Those hours don’t come back, and morale drops when effort doesn’t turn into conversations.
The second is your domain’s reputation. With badly segmented lists, bounces, unsubscribes and spam flags go up, and email providers take note. There comes a point where even your well-targeted messages end up in the junk folder.
The third is a pipeline that looks healthy but isn’t. Meetings with companies that have no budget or no decision-making power inflate the funnel and then stall before the proposal, which distorts your forecasts.
Validating first with contained volume limits all three costs. If a segment fails, you have only spent a small part of your market and your reputation.
How to set up tests by segment
You can’t test a whole ICP at once, because if the results are weak you won’t know which criterion fails. The solution is to split it into segments and test them separately.
Split the ICP into comparable segments
A segment is a group of accounts that share the variable you want to test. For example, you can split by industry, size, region or timing signal (companies hiring in sales, that have just opened a market or have changed CRM). Choose one variable per test. If you mix industry and size in the same test, you won’t know which of the two explains the result.
Change only one thing at a time
For the comparison to be fair, everything else has to be the same across segments: the same sequence, the same number of touches, the same cadence, the same type of sender and the same offer. The only thing that changes is who you write to. If you also change the copy, you are no longer validating the ICP, you are validating the message. If you need to build that common sequence, here we explain how to do lead follow-up without being annoying.
Turn each segment into search filters
“Mid-sized tech companies” is not a filter. “B2B SaaS companies in Spain, of a size that fits your price, with a sales job opening published in the last few weeks” is. Each segment must be reproducible in your search tool, so you can pull the same list again when you decide to scale it. Check contact data before each batch: people change company and role, and an old database contaminates the test.
Want to put this to work this week? Grab B2B prospecting guides and checklists to go from theory to your next sequence. Get the SDR Pack
Sample size: how many contacts per segment
Sample size can go wrong on both sides. With too few contacts, a single reply changes the rate abruptly and any conclusion is chance. With too many, you are no longer testing: you are scaling without having validated.
The logic for deciding is this:
- The lower you expect the rate to be, the more contacts you need. If the events you want to measure (positive replies, meetings) are infrequent, you need more sends to see enough of them to compare.
- The smaller the difference you want to detect between segments, the more contacts you need. Telling apart two very different segments is easy; telling apart two similar ones takes a much larger sample.
- Fix the sample before launching, not during. If you stop the test as soon as one segment “seems” to win, you are most likely looking at noise.
- Respect the real size of the segment. If the whole segment is small, don’t spend most of it on the test. In that case, validate with fewer contacts and lean more on the quality of the replies.
Define your minimum sample threshold per segment using your own historical data and keep it the same across all segments in the same test. What matters is that the rule is the same for all and is written down before you start.
What to measure: reply rate and meeting rate by segment
With the sample defined, measure each segment separately. Two metrics cover the basics:
- Reply rate = replies received ÷ contacts who received the sequence.
- Meeting rate = meetings booked ÷ contacts who received the sequence.
Neither is enough on its own. A high reply rate with many “not interested” or “I’m not the right person” replies shows that the message catches attention but the segment doesn’t fit. That is why it pays to classify replies into positive, negative, referrals to someone else and unsubscribes, and to also calculate the share of positive replies out of the total.
Meeting rate is the closest to what matters to you, but it doesn’t settle the question either. If you can, follow each meeting one step further: does it move on to a proposal? A segment that books meetings that never get past the first call has surface-level interest, but lacks need, budget or decision-making power.
Also read what the replies say, not just how many there are. Repeated phrases such as “doesn’t apply in our case” or “we already solved it another way” tell you which ICP criterion fails. To go deeper into how ICP and ECP fit together when reading this data, the ebook on ICP and ECP covers both profiles in more detail.
Criteria to scale, adjust or drop a segment
Before launching, write down which result will lead you to each decision. If you decide after seeing the data, you will tend to justify what you already wanted to do. A simple way is to set two thresholds per metric: one above which the segment deserves volume and one below which it doesn’t deserve more effort. Define your threshold based on what you need for the channel to be profitable: how many meetings you need to close a sale and how much each contact costs you.
Scale
The segment beats your meeting-rate threshold, positive replies outnumber negative ones and meetings move on to a proposal. In that case, raise the volume gradually and keep measuring, because as the list grows, accounts that fit worse than the first ones come in.
Adjust
The segment falls between the two thresholds, or one metric is doing well and another isn’t. This is the most common case. Look at where the failure is: if the wrong people reply, review the roles; if they reply well but don’t move forward, review size, budget or timing signals; if there are unsubscribes, review the exclusions. Change one criterion and test again with a new sample.
Drop
The segment falls below your lower threshold and the replies confirm there is no fit. Remove it from the ICP or move it to exclusions, and note why. Dropping is also a result: it saves you the cost of scaling something that wasn’t going to work.
When you spot a weak segment, the temptation is to change the copy. Resist it until you have reviewed the recipient. Many times the message was right and the segment wasn’t.
Checklist before raising volume
Before hiring more SDRs, outsourcing the work to an agency or multiplying sends, go through this list. If any box is empty, it isn’t time to scale yet.
- ICP written down and split into segments with a single variable per test.
- Same sequence and offer across all segments in the test.
- Minimum sample per segment fixed before launching.
- Scale and drop thresholds written before seeing the data.
- Reply rate and meeting rate calculated per segment.
- Replies classified as positive, negative, referrals and unsubscribes.
- Meetings followed until you know whether they move on to a proposal.
- Decision noted for each segment: scale, adjust or drop.
With all of them ticked, volume stops being a bet and becomes a lever. At HelloMrLead we work in this order because booking meetings with people in your target depends more on choosing the right account than on sending more messages. You’ll find more practical material in our resources for SDRs.
The next time someone proposes “doing more” to fix weak results, ask first for the data by segment. With that answer in hand, any new sequence has a better chance.
Move from document to action. Grab B2B prospecting guides and checklists to go from theory to your next sequence. Get the SDR Pack
