channel metrics that tell you something: sourced, influenced and the partner health score

If most of your revenue goes through partners, “partner revenue” is just revenue. The useful questions are who created the demand, who helped win the deal and who fulfilled it. Record the origin of every deal once, at creation, and lock it. Report sourced and influenced revenue side by side, never added together. And sort your partners with a health score that weights leading indicators at least as heavily as last year’s revenue.

the most common metric is the most incomplete

In the 2022 State of Partner Ops and Programs survey by HubSpot, Canalys and Partnership Leaders, 81.1% of organisations used partner-sourced revenue as a partner KPI, against 58% for influenced revenue. Sourced revenue is easy to defend, because a registration proves it. It also undervalues partners that do the work on deals they did not find: the integrator who ran the POC, the MSSP whose analysts convinced the customer.

Forrester argued in 2025 that attribution focused on sourced revenue misses partner influence across the buying journey, and that company leadership has to define partner classifications and what each is worth.

three questions, three fields

1. originwho created it?Partner, your marketing, your sellers, the customer. Set at creation, locked after the registration decision.
2. influencewho helped win it?Any partner with a logged contribution: POC, integration, sponsor access. Can be several.
3. fulfilmentwho transacted it?The reseller or distributor of record. Drives rebates and tier revenue.

Rules that keep it clean:

1. Origin is the primary attribution. Only partner operations can change it, with a logged reason.
2. Fulfilment drives rebates, tier credit and revenue reporting.
3. Influence is recorded as a partner linked to the deal with a role, plus evidence. A workshop held counts. “We know the customer” does not.
4. A report of influenced revenue counts each deal once, however many partners touched it.
5. Publish the attribution rules in your partner guide, so partners can predict their tier credit.

be careful with partner-lift numbers

The direction of the published evidence is consistent: deals with partners involved tend to win more often and be larger. The size of the effect is not. Crossbeam reports an average 11.7% win-rate lift from its network data. A Canalys study cited by AWS found that 65% of frequent co-selling partners reported higher close rates, which is a share of partners reporting an improvement, not the size of it. All of these come from companies that sell partner tooling or programmes, and all share a bias: partners get invited into deals that already look promising.

Measure your own uplift instead. Take closed deals, won and lost, from the last four to eight quarters. Group them by product, deal size band and segment. Within each group, compare deals with a partner in an influencing role against the rest, on win rate, deal size, cycle length and discount. Show percentages only where each group has 20 to 30 closed deals or more. Below that, show the raw counts.

The most useful comparisons tend to be certified against non-certified partners, deals where your engineers were involved against deals without them, and partners with a joint business plan against those without. Those tell you where your enablement and presales time pays back.

define “active” before anyone uses it

“Active partner” means five different things in five meetings unless you fix it. Use four states, each with a written test.

State	Test	Common trap
Activated	First customer-facing work in its role, accepted by the owner. A date that never changes	Counting a course, a login or a promise
Active	Documented customer-facing activity in the last 90 days	Marking a services partner dormant because it had no new logos
Commercially active	A qualified opportunity with a current next step, or a transaction in the window	Counting a portal login
Productive	Real financial contribution over 12 months, above a threshold for its segment	Treating a demo licence order as productivity

The 90-day and 12-month windows are starting points. Pick yours and keep them for at least a year, or your trend lines mean nothing.

a health score you can build in a week

A partner health score puts each partner’s state in one number with visible parts, so a channel manager can sort the portfolio on Monday morning.

performance: revenue against target, growth25%

pipeline: registrations, qualified pipeline25%

capability: certified people, demo environment15%

engagement: QBR attendance, responsiveness15%

customer outcomes: renewals, escalations10%

hygiene: forecast reliability, data quality10%

Add a floor in each dimension. Microsoft’s partner capability score is a public example: it sets a 70-point threshold and requires every metric to be above zero, so a partner can’t reach the bar on revenue alone with no certified people. And treat a missing input as unknown, never as green.

one dashboard per audience

Audience	Question	What it shows
Leadership	Are we on plan, and where is the risk?	Revenue against plan by product, forecast, top partner concentration
Channel lead	Which partners need attention?	Pipeline coverage, health score distribution, registration ageing
Channel manager	What do I do today?	My partners’ scores, pending registrations, renewals in 90 days
Presales lead	Is engineer time going to the right deals?	Engineer hours by deal and stage, POC outcomes
Partner	How are we doing together?	Its revenue against the joint plan, pipeline, certifications, renewals due

Keep each dashboard to 8 to 12 metrics, open every one with the comparison to target, show at least four quarters of trend, and never show a partner another partner’s name or numbers.

data quality is the real limit

Gartner reported in 2020 that fewer than half of sales leaders and sellers had high confidence in their forecasting accuracy. Channel data is worse, because it arrives from people who don’t work for you. Mandatory fields, validation rules, one named data owner and a monthly audit do more than another dashboard. Remove any metric nobody has acted on for two quarters.

the question for your channel team

If you removed every partner-sourced number from your next board report, what would you still be able to say about your channel?