Customer discovery interviews, and how many are enough
B2B customer discovery interviews: the sample size the research converges on, why your segments change that number, and the questions that get honest answers.
Read
BDGL / Insights / Partnerships
A partner who has learned your product and a partner who has closed one deal with it are in different states, not at different points on one path. The vendors publish which one they count.
Most partner programmes have a signature problem. A reseller signs, gets a welcome pack, sits through a product session, receives a portal login, and is never heard from again. Nobody can point to the moment it failed, because nothing visibly failed. The paperwork all completed.
What failed is that onboarding was designed to transfer information when it needed to produce a transaction. A partner who has learned your product and a partner who has closed one deal with it are not at different points on the same path. They are in different states, and only the second one is actually onboarded.
The useful thing is that the big vendors publish exactly what they count, in numbers. You can read their thresholds and see what a serious programme expects a new partner to reach. Whether you are building a programme or joining one, those published numbers are the clearest available answer to the question "what should the first ninety days produce".
Take the AWS services partner tiers, which set out what a consulting partner must hold at each level. The entry tier is not a badge for attending a webinar.
| Requirement | Select | Premier |
|---|---|---|
| Accredited individuals | 4 (2 technical, 2 business) | 20 (10 technical, 10 business) |
| Technical certified individuals | 2 | 25, min 10 Professional or Specialty |
| Launched opportunities | 3 | 50 |
| Total monthly recurring revenue | 1,500 USD | 50,000 USD |
| Annual programme fee | 2,500 USD | 2,500 USD |
Three things in that table change how you would run onboarding.
The first is that the entry tier requires launched opportunities, not training completions. Three of them, carrying at least 1,500 US dollars of combined monthly recurring revenue. A partner cannot reach the lowest tier by learning. They reach it by selling something that goes live.
The second is that the fee is flat across tiers. Select, Advanced and Premier all cost 2,500 US dollars a year. So the programme charges the same whether a partner produces three launched opportunities or fifty. A partner sitting at zero is paying the same as one at the top, which is a quietly powerful argument to use in a stalled onboarding conversation.
The third is at the top end. Premier requires sustained attainment of the tier criteria for more than six months, plus a partner business plan and an executive business review. The programme is explicitly not satisfied by hitting a number once.
Microsoft's version makes a sharper point. Its Solutions Partner designation is scored out of 100 points per solution area, built from three categories: performance, skilling and customer success. To qualify, a partner needs at least 70 points, and every metric in the solution area greater than zero points.
That second condition is the one worth pinning to the wall. A partner can accumulate a high total and still fail because one metric sits at zero. Microsoft's own documentation describes the "in progress" status as what you see when you have earned some points but either have less than 70 total, or have zero points in an individual metric. The partner capability score documentation sets out the categories and the qualification rule.
Translated into onboarding terms: a partner who has certified five engineers and sold nothing is not 70 percent of the way there. They are failing, and the scoreboard says so plainly. Skilling without a customer is a zero in two of three categories.
There is a detail in the same documentation worth knowing if you run a programme on someone else's platform. Qualification only has to be met on one day inside a window that covers the current month and the previous five full calendar months. So the partner's status is a rolling snapshot, not an annual verdict, and a partner can slip out of qualification without any single event causing it.
Working backwards from those thresholds, the sequence is not the one most programmes run.
The ordering is the whole argument. Most programmes do certification first because it is easy to administer and produces a number quickly. It is also the step a partner is least motivated to complete before they have seen any money.
Someone on your side has to own the partner's first transaction by name, with it on their own number. Not "supporting the channel". Owning that deal.
This is where co-selling arrangements quietly fail, because the answer to "whose customer is this" was never agreed. Decide before the joint call who holds the relationship if it closes, who holds it if it stalls, and what happens if your direct team is already in the account. The clauses that decide this belong in the agreement rather than in an email, and we set out which ones matter in channel partner agreement essentials.
If the model itself is still open, the question of who owns the end customer differs sharply between reselling and distribution, which we compared in reseller versus distributor model.
Everything above is a checklist with an owner and a record, which means it survives only if it lives somewhere other than one person's memory. Partner onboarding goes stale exactly the way employee onboarding does, and for the same reason: the steps are known, nobody disputes them, and they still get skipped because no system is asking.
The engineering view of that problem is worth reading even if you never build the tool. The team at Rivl wrote up what an internal onboarding tool actually has to hold, and the structure transfers directly: a checklist, an owner per step, a record of what happened, and a visible state per person. Swap "new hire" for "new partner" and the design is the same.
Four failures account for most of it.
Recruiting faster than you can onboard. Twenty signed partners with no first deal is worse than four producing, because the twenty consume attention and generate reporting that looks like progress. If you are still adding partners, the sourcing routes are in how to find channel partners in the Gulf, but fix throughput first.
Treating the portal as the programme. A login is not enablement. Nobody has ever been activated by a document library.
No margin clarity before the first quote. A partner who cannot calculate what they make will not put your product in front of a customer, and will never tell you that is the reason.
Choosing the wrong partnership model for the effort available. A referral arrangement and a reselling arrangement need very different onboarding, and the three models ranked by effort are in B2B partnerships strategy. If you want the lightest version that still produces something, how to build a referral partner program is the one to start from.
Three caveats, because the numbers above are precise about the wrong thing if you read them carelessly.
The AWS and Microsoft thresholds describe what those two vendors require of their partners. They are not benchmarks for your programme, and copying 1,500 US dollars of monthly recurring revenue into your own tier structure because Amazon uses it would be borrowing a number from a business with nothing in common with yours. Use them as evidence that serious programmes define a measurable finish line, then set your own.
Published programme requirements also change, and both sets carry their own caveats. Microsoft notes that required certifications are expected to change as solution areas evolve, and that data refreshes on a monthly cycle, so a partner's score is never quite live. Check the current pages before you quote a figure to a partner.
And there is no published figure here for how many partners never sell after signing, because we do not have one worth citing. The pattern is what we see in the programmes we work on, which is an observation rather than a statistic. If someone quotes you a precise industry percentage for partner inactivity, ask where it came from before you put it in a board pack.
Send us how many partners you have signed, how many have launched a deal, and what your onboarding currently asks them to do first. We will tell you which step is out of order and which partners are worth the next ninety days.
Book a Free Consultation