Most partner programmes are four programmes pretending to be one.
Different economics. Different enablement. Different deal cycles. Different reporting lines. Different KPIs. Different problem when something breaks.
The board pack does not show any of this. It shows one line called partners, one attach rate, one revenue number, one head count. So the board sees a portfolio. What you actually have is an architecture, and each part of it is loading a different beam.
I have run programmes that looked healthy on the board slide and were falling apart underneath because the reseller economics were carrying implementation work that should have been sold direct. I have watched CROs discover that their MSP tier had eaten their SI tier, because both got the same comp plan and the cheaper motion won. I have watched founders defend a technology alliance track that never contributed a customer, because the logo mattered more than the money.
Every one of these failures is the same failure. Four programmes running under one label, one owner, one measurement system.
The fix is not to merge them. The fix is to separate them, name each one, and give each one the ownership, compensation, enablement and board reporting it actually needs.
That is the architecture.
Ortent published a diagnostic on this earlier in the summer. It runs a twelve-marker A/B assessment across three components. Ownership and reporting. Enablement and compensation. Board-facing signals. It measures how far your programme has blurred the four archetypes together: reseller, systems integrator, ISV, and instrument or platform vendor. Every other partner type the industry describes is a variation on one of these four.
The diagnostic ends with a scored read: how much of the programme is blurred, which of the three components carries most of it, and where to start.
Two ways to use it. Run it yourself in ten minutes and see what you get. Or run it with your CRO in the room, which is where it earns its keep.
The paper and a visualiser of the four archetypes are at ortent.co/tools/partner-architecture. The twelve-marker diagnostic itself runs as a Claude prompt: it takes you through one marker at a time and scores the programme at the end.
If you run a partner programme, or sit on the board of a company that does, this is worth thirty minutes.
One warning. The diagnostic does not lie. If the programme is three archetypes short of the one on the deck, you will see that. Whether you want to see it is a separate question.