RU
ICP TEST

Test a B2B ICP Before You Hire Sales

Test a B2B hypothesis before hiring sales by running a limited outreach campaign across a few distinct segments. It shows which companies recognise the problem, which trigger earns a reply, and where the hypothesis breaks down. This is market validation before building an SDR process, not a substitute for sales.

Case studies
5.93%
of contacted companies replied across our campaigns
6,500
clear yeses in our own campaigns

How to tell whether the issue is ICP, not sales execution

A common conclusion is that you need a stronger salesperson. But if that person receives a list of companies without a clear reason to contact them, they must discover the market, explain the value and close the deal at the same time. You cannot interpret the result: a rejection may mean the company is wrong, the offer is wrong or the conversation was weak. Before hiring, separate market demand from sales skill.

The signal is not a lack of deals after a week, but conflicting replies. An equipment supplier may write to plant directors about lowering production costs and receive requests for a catalogue, referrals to the chief engineer and statements that nothing is planned. Those are different situations: the first points to a person and procurement context, the second changes the route, and the third suggests the timing is wrong.

Do not test an abstract profile such as a company of a certain size. Test a combination of company type, decision-maker role, an observable sign of the task and a specific next step.

You can check the terminology in the <a href="/en/glossary/">outreach glossary</a>, but write the working hypothesis as one sentence. For example: write to an operations director at a regional distributor that has opened a new warehouse and offer to compare its picking process against two operational indicators. If you cannot write that sentence, it is too early to scale sales.

The three versions that make up an ICP test

Do not test one broad audience and conclude that the market does not respond. Build three versions that differ by one material signal. For an automation service, these may be companies hiring operators, companies with multiple branches, and companies whose job listings mention several business systems. Record the public evidence beside every company: its website, vacancy, directory listing, procurement notice, registry entry or news item.

VersionWhat you can see before writingFirst question
Branch networkThe website lists several citiesWho brings branch processes into one system?
Growing teamAn operator vacancy is openHas workload already increased, or are you hiring ahead of demand?
Changing systemsA vacancy names two systemsIs the task integration now, or manual reconciliation?

Do not combine these versions in one email. “We help businesses automate processes and reduce costs. Interested in a conversation?” gives the recipient no reason to reply. A more useful approach is to name the observed signal and ask one narrow question. It does not promise savings without evidence, and it leaves room for a clear “not relevant.”

Do not take contacts and roles from an unfiltered export. An ICP test needs companies with a verifiable signal and the person who owns the process. See how this is handled in <a href="/en/services/lead-list-building/">B2B lead list building</a>.

How to run a test without pretending it is a large funnel

  1. State the product hypothesis: which costly or recurring problem the product solves, and what the customer should confirm in a reply.
  2. Choose three segments and define inclusion criteria, exclusions and the recipient role for each.
  3. Prepare a distinct opening line and question for each segment, changing one meaningful element at a time.
  4. Send the initial touches through prepared infrastructure, record replies verbatim and classify the reason: interest, another owner, current supplier, poor timing or an unsuitable profile.
  5. Review the result as a map of segment, trigger, role, reply wording and next step, rather than one percentage.

Do not confuse a technical issue with a product test. A nonexistent address is not an ICP rejection. In OT9’s own campaigns, 97.7% of addresses accepted the email and 2.3% bounced. Verify addresses before judging a segment and, where appropriate, look for a second relevant contact; see <a href="/en/services/list-verification/">email list verification</a>.

Do not automatically count “send a proposal” as validation. Ask what the proposal should address: reducing manual work, controlling branches or something else. When the recipient names a situation, role or timeframe, you have material to qualify; collecting documents without context is not yet a reason to scale.

What you can learn in the first weeks

You can draw an initial conclusion once recipients have had time to read and answer the first touches, but it does not predict a contract date. Across our campaigns, 6,500 companies have given a clear yes by asking for details, a proposal or a meeting. That shows that specific interest can be separated from noise; it does not set a benchmark for your product, price or audience.

A positive test is specific: several recipients in one segment name the same problem, route you to the right role or ask to discuss their case. A negative test is useful too: companies reply but do not recognise the situation, the role repeatedly redirects the message, or the question only produces requests for a generic price list. Change the hypothesis, the signal, task owner or offer wording—not simply the list size.

  • Every company is in the segment for a stated reason, not only its industry code or region.
  • For every reply, you know what was confirmed: role, problem, timing, current process or willingness to meet.
  • Rejections are grouped by reason instead of being filed as one generic “not interested.”
  • The sales team receives a repeatable combination of segment, trigger and first question, not only a contact list.

Once the combination repeats, you can move it into an ongoing outbound motion or <a href="/en/services/full-service-outreach/">full-service outreach</a>. Before then, expanding the list usually produces more ambiguous replies.

When outreach will not give you an honest answer

Outreach is not the right test if the product cannot be explained without a long diagnostic process and you do not have a specialist ready to run that conversation after a reply. It also does not replace research when buyers do not yet recognise the category: they may confirm the problem exists but lack a way to evaluate a new solution.

Do not start if you do not have the right to sell the product, cannot state even a price range, or cannot accept a meeting and answer questions promptly. Getting an interested reply and then spending a week deciding what you can offer tests internal readiness, not ICP.

If you need to speak with a small set of known strategic accounts, use a focused <a href="/en/services/account-based-marketing/">account-based marketing</a> approach instead. If you already have a list but have not analysed it, begin with an <a href="/en/services/outreach-audit/">outreach audit</a>.

FAQ

Can you test an ICP before the product is on the market?

Yes, if you can show a credible use case and hold a substantive conversation after a reply. Do not pretend you have a finished case study; test whether the described task exists and whether the recipient is willing to discuss solving it.

How many segments should the first test include?

Three distinct versions are usually enough. More versions blur the conclusion because role, trigger, wording and company selection start changing at once.

What counts as confirmation of the hypothesis?

A reply alone does not confirm it. Look for a repeatable, specific signal: the recipient recognises the situation, names the process owner, asks to discuss their case or agrees to a conversation. A request for a generic proposal needs further qualification.

Should you hire SDRs immediately after the test?

No. First create a transferable instruction: whom to contact, which observable sign to use, what question to ask and how to handle common replies. Otherwise, each SDR will begin the research again from zero.

Can you test ICP through Telegram or LinkedIn?

Yes, when the required role is genuinely reachable in that channel. The logic remains the same: one segment, an observable trigger, a short question and analysis of what replies say—not only how many arrive.

Build an ICP test before you hire

We will review your product, target-profile versions and list signals. You will know what to test in the first touches and which replies should count as evidence.

Pricing
24 hours
that is how long it takes us to come back with numbers for your segment