The required internal process beneath the partner journey. Every task has a named owner, agreed actions on both sides, a data signal it must emit and an answer to what breaks without it. The task list will grow and shrink as the pilot teaches us. The discipline will not.
Download the full matrix (Excel)How to read this page: each card shows the 4 load bearing answers. The spreadsheet holds the full detail, including track variations, tooling fields and the deadlines on both sides. Owners: PEM is the partner enablement manager, PDM the partner development manager, Partner is theirs.
Before any kickoff. The quality of everything downstream is decided here, which is why cold volume recruiting is the first thing this model refuses to do.
Census of existing partner-adjacent work. Candidates scored on capability, live customer conversations and evolution potential.
Discloses real pipeline. Names an executive sponsor.
Candidate score and source logged.
Cold recruits become the dormant 40 to 60%.
Issues a short questionnaire. The key question: name your top 3 live customer conversations. The answer sets the support level.
Answers straight. 3 named opportunities, or the real number.
Completed questionnaire, count of named opportunities.
Kickoff goes generic and support gets spent blind.
Agreement executed, portal accounts live, deal-registration access provisioned. All admin executed before anyone meets.
Signs, nominates users, completes first login.
Provisioning date. The SLA clock starts here.
Week 1 becomes form filling and the momentum never arrives.
Five working days that decide whether this partner ever activates. Every hour spent here is spent on a real deal, never on ceremony.
Runs a 90 minute kickoff. The anchor deal is chosen together from the questionnaire's named opportunities.
Sales lead and executive sponsor attend. Commits to the anchor deal.
Kickoff date. Opportunity named, yes or no.
No anchor deal, and the silence starts immediately.
Screen share, their hands on the keyboard. A real registration if legal is done, a practice run if not.
Their seller performs the registration personally. Not their admin, not us.
First registration event, real or practice.
Untaught registration means an empty portal at day 30.
An ADI engineer works the technical fit, parts and reference designs, hours drawn from the cohort budget agreed before launch.
Their engineer and seller attend, prepared, with the customer problem defined.
Session date. If unbooked by day 10, it escalates inside ADI, not onto the partner.
Opportunity #1 stalls untouched and week 1 was theatre.
Paths by role: engineer, seller and exec. Sandbox credentials on the software track. Nobody gets a generic library.
Named people start their paths inside seven days.
Path started, percentage within 7 days.
Generic training gets ignored, and certification never starts.
30 minutes. Blockers named, next registration targeted, the partner's score switched on.
Their seller attends and names blockers plainly.
Checkpoint held. Blocker log opened.
The first fortnight drifts silently.
The stretch where intent becomes evidence. It ends at the model's one hard gate.
Reviews against the quality bar within 48 hours: named customer, named application, real timeline.
Registers only real opportunities. The bar exists to protect the field's trust.
Time to first qualified registration. The programme's north-star number.
Junk pipeline costs the field's trust, and trust does not come back.
The early warning score runs from day 7. Defined triggers fire defined actions with named owners.
Nothing extra. The score reads signals that real work emits anyway.
Score, plus every trigger event with owner and action.
Disengagement gets noticed at day 60, which is too late.
Gate review. A miss fires a diagnosis call, never a warning letter. Cause recorded accurately. Governed exceptions carry a named reason.
The registration, or a working proof on the technical track, or the real reason.
Activation yes or no, days to gate, miss-cause code.
Zombies corrupt every number downstream.
One deal could be luck. This phase exists to prove there is a repeatable motion.
Supports from the second chair. The partner leads.
Fronts a real customer conversation on the anchor deal.
Meeting date and a short debrief note.
The programme stays theoretical.
Reviews against the same quality bar. Coaches where the pipeline is thin.
Registers the next 2 real opportunities.
Registration count and velocity.
A single-deal partner is a project, not a partnership.
Tracks completion by role. Certification gates tier benefits, never the right to sell.
Named people at roughly 50% of their paths.
Completion percentage by role.
Selling without depth, and the engineering dependency never falls.
The scaling economics only work if partners learn to run without us in the room.
Stays out of the room, deliberately. Debriefs after.
Runs a customer conversation on ADI solutions with no ADI present.
Independent meeting logged.
Permanent dependence kills the scaling economics.
Forces the call. A deal closed with a reason is a healthy signal.
Moves the stage, or closes it and says why.
Stage change, or disqualification with reason.
Polite fiction accumulates and reviews become theatre.
Evidence pack and a recommendation. Contested cases go up a level for decision.
Attends and commits to the outcome: graduate, extend with a named fix or move down a tier.
Exit status. Cohort graduation rate.
Support gets spent where it returns nothing.
Graduation is the start of the relationship, not the end of the process.
Runs the QBR: pipeline, enablement gaps, next quarter's joint plan.
Brings the numbers and the asks.
QBR held, actions logged.
Post-graduation drift, and invisible year-two attrition.
Reviews earned progression. Switches tracks on as the partner's shape changes.
Earns it: certifications, registrations, joint wins.
Tier changes. Tracks active per partner.
The evolution engine never runs, and hardware partners never become solution partners.
3 questions to every active partner monthly. One process change published per month. At month 6, the score weights are tested again against who actually reached design-in, and changed.
Answers 3 questions straight.
Friction log. Weight-change record.
The machine ossifies, and version-one judgment hardens into dogma.