The FY27 co-op guidelines run to around 300 pages. Nobody's reading that cover to cover, and Copilot will summarise it happily but miss the bits that cost you money.
So Molly and I ran a session to strip it back: what's changed, what it means in real terms, and what good actually looks like when you spend the money.
Here's the short version.
This isn't a year of tweaks
In a normal year, Molly's advice would be to check your activity names and carry on. Not this year.
If you're working from a plan that was signed off in FY26 and you've carried it forward, it's not that it's just a little bit out of date - parts of it are now basically completely not fundable, and you probably won't find out until the point that you go to claim for it.
Molly Ridley, Head of Partner Growth & Marketing, Noteworthy SupportThe big structural change: Microsoft has pulled apart the single commerce incentives guide and replaced it with three scenario-based investment programmes - AI Business Solutions, Cloud and AI Platforms, and Security. The bookmark your team has been using for years is out of date, and before you can work out whether an activity is eligible you need to know which programme it sits under.
Three knock-on effects worth knowing
Fewer generic opportunities
The catch-all earning activities that used to absorb anything broadly marketing-shaped have narrowed. A plan written in general terms now fails earlier, which is better than failing at claim stage, but vague plans get bounced.
A harder focus on AI and growth
Microsoft wants this money moving adoption, not topping up a marketing budget you'd have spent anyway.
A stricter bar on proving delivery
This is the one Molly flagged hardest, because it's what turns a perfectly good approved plan into a rejected claim months later.
The names have changed, and that matters more than it sounds
Direct mail, email, SMS, digital campaign and social have all been rolled into one activity: paid media with multi-touch campaign. Trade shows are now events, customer and conferences. Telemarketing is targeted prospecting. Customer seminars are customer workshops. Migration services sit under customer solution adoption.
Renaming sounds trivial. It isn't, for two reasons.
The admin one: the activity you select has to match the activity you actually ran, and mismatches are a very common reason for claims coming back for clarification. Clarification costs weeks, and sometimes those weeks cross a deadline.
The bigger one: consolidation changes the requirements. Paid media with multi-touch isn't five old things with a new label, it brings a whole new set of conditions. Paid media is now non-negotiable within it, and you need at least two other supporting activities alongside it.
Which is less scary than it sounds. If you ran ads, you almost certainly sent people to a landing page, and that's component one. Optimise that page for SEO and AEO and you've got component two. You don't need to create ten thousand things. You need to do a couple of things properly, and evidence them.
Practical action: take your current plan, map every line to the new FY27 names, and flag anything that doesn't map cleanly. Whatever doesn't map cleanly is where your eligibility problem is. There's more detail in our Microsoft Co-op funds hub.
What's gone, and what's capped
Microsoft syndicated content, proof of concept and employee purchase web setup are no longer claimable under any programme, with no alternative. Print and postage are out full stop, which is a genuine shame because I'm a big believer in direct mail. Use your own budget for the print and postage, keep it in the multi-touch mix, and fund everything else with co-op.
Also out this year: entertainment venues, celebrity and motivational speakers, drinks reception sponsorship, advertising on your own website or on-hold phone system, and Partner of the Year nominations and their assets. Branded merchandise is allowed but capped at 1%, with the Microsoft or product name in text only. As Molly put it: "you can fund the golf balls, but not the artwork."
On caps for CSP and hosting partners: allocated resource is capped at 50%, partner AI adoption and customer solution adoption at 5% each, marketing automation at 3%, internal incentives and SPIFs at 2% each. Those percentages are of your co-op allocation, not your overall marketing budget, so run the maths against your real figure.
Two things to plan for. The 50% allocated resource cap is expected to fall again in FY28, and my gut says towards 30%, so if your model leans heavily on people's time, start planning for that now. And if any non-Microsoft product is present in the activity, your claim is capped at 30% of the activity cost. Know that before you commit the spend, not after.
Marketing resources also need to complete the new Microsoft GTM Titan training, with three months from August. Sales resources need their own aligned training in the same window. We're already in September.
Where Microsoft wants the money to go
The guidance splits by solution area: 30% Cloud and AI Platform, 30% Copilot, 20% Security, 15% AI workforce excluding Copilot, and 5% AI business process.
It's guidance, not a hard rule, and nobody's rejecting a claim because it's 40% security. But it's genuinely useful intelligence, because it tells you what Microsoft is being measured on this year. That shapes which conversations your Microsoft contacts want to have and where co-sell discussions go more easily.
Lay your planned activity over those five buckets and see what falls out. A lot of partners find they're heavily weighted towards whatever they sold well two or three years ago. Much better to have that conversation in September than in June with the window closing.
The dates - because this is where partners lose real money
Co-op earned in FY26 H2 (Jan-Jun 2026) is spent Jul-Dec 2026 and claimed by 15 February 2027.
Co-op earned in FY27 H1 (Jul-Dec 2026) is spent Jan-Jun 2027 and claimed by 15 August 2027.
Claims open 45 days after the usage period begins, so there's a stretch where the activity has happened but you can't submit yet, and that gap is where things get forgotten.
Nothing rolls over. Nothing is half paid. Late doesn't carry into the next period, it's forfeited in full.
Put both deadlines in a shared calendar today with a reminder a month before each, and treat the deadline as the date the claim and its evidence are finished, not the date you start assembling them. We also do an A6 desk calendar with the co-op dates on it, so say the word and we'll send you one for next year.
Compliant and well spent are two different things
That's the rules. Here's the bit I care about more.
There's a big gap between spending money compliantly and spending it well, and a lot of partners land firmly in the first camp. Co-op gets treated as a shopping list. Nine activities at a few hundred or a couple of thousand pounds each, four audience types, three propositions: AI, security, modern work. Come January you've got nothing to show for it. Nothing reusable, nothing running, no data.
The other version: one audience, one proposition, six joined-up activities across a half. A webinar, some blog content, paid media doing the heavy lifting, a proper nurture path. By month seven you've got engaged prospects, an asset library, case studies, and data that tells you what to do next. Pick the financial services blueprint up, move it to retail, run it again.
Both are claimable. Only one moves the business forward.
Microsoft's own numbers back this up. Around 70% of partners underinvest in demand generation, and 43% use less than half the funding vendors make available to them. Investing in high-ROI activity can drive two to five times greater return than merch, a golf day or amplifying a blog. Concentration pays.
The other reason to care: Microsoft is rewarding growth now, not retention. Spend co-op badly enough, for long enough, and the likely outcome is less co-op to spend.
What a 12-month programme looks like
The mindset shift I'd most like partners to make is treating co-op as a 12-month programme rather than a 30, 45 or 60-day sprint because there's budget about to expire. I've seen partners with anything from $14,000 to $140,000 and six weeks to spend it. Spending it right and spending it quickly aren't the same thing.
The five layers we talked through
- 1
Paid media
The anchor. Ad spend, placement, database enrichment, Sales Navigator.
- 2
Trust and credibility
Traffic's useless if people land somewhere that gives them no confidence. Cost per click and cost per acquisition are the highest we've seen in a long time, so the landing experience has to earn it.
- 3
The conversion point
A webinar, an assessment, a workshop, a demo. Something tangible.
- 4
Nurture
The forgotten layer. They came to the webinar, then went quiet. What happens next? This is where sales and marketing working together really shows.
- 5
Longevity
Case studies that stand the test of time, content you can repurpose, data that informs the next campaign, a blueprint you can lift into a new vertical.
The thread running through all of it: build something with legs. Rinse and repeat beats starting from scratch every single time, and we do this for a living, so trust me, starting from scratch is the hard way.
One more thing I'd push you on: first-party insight. Recurring customer questions and your most-read blogs are free and useful, but every other partner is talking about the same things. Run a survey into a specific customer set and you get data nobody else can quote. It's brilliant for credibility, it does real work for your SEO and AEO, and it means you're saying something objective rather than "well, we've had two conversations about AI governance this week, so everyone must care about AI governance."
Getting the claim approved
You can do everything right on the activity side - right bucket, right solution area, inside your caps, beautifully executed - and still lose the money because the claim's wrong. And at that point there's nothing anyone can do. The activity's happened, the invoice is paid, the reimbursement's gone.
Remember it's a reimbursement model. You fund it first and get the money back at least 60 days after Microsoft's assessment, so finance needs to understand the shape of that, especially if you're spending a big chunk near the end of the window.
What Microsoft wants to see
- A screenshot of the ad as it ran
- The creative you used
- Platform reporting showing impressions, clicks and leads
- A detailed third-party invoice
- One line item per activity on the CSR form
The invoice is where most rejections happen. "An invoice that says 'campaign services' and gives one figure isn't evidence, it's just a summary." You need a breakdown clear enough that someone with no context can read it and understand exactly what was bought, for which activity. Ask for that at the point of invoicing, because asking three months later, when the account manager's changed, is a different job entirely.
And use the creative pre-approval process. Microsoft typically turns it around in three business days, and it's genuinely underused. Three days of waiting against the risk of a rejected claim months later is a very good trade, especially the first time you run a new activity type.
Five things to take away
Collect evidence from day one
Not at claim time. Decide who owns the evidence file when you plan the activity, and put the requirement in your agency brief up front.
Build paid media in from the outset
Every multi-touch campaign needs it, so don't bolt it on when someone realises it's mandatory.
Check your spend against the new caps
Do it before you submit, especially if you're a CSP or hosting partner.
Track the deadlines
It's all or nothing. Nothing rolls over and nothing is half paid.
Avoid the prohibited costs
Travel, meals, print and postage, event entertainment. And watch for partial duplicate claims, which almost always happen because two people are assembling two claims and neither knows about the other.
The partners who consistently get the most out of co-op aren't necessarily the ones with the smartest campaigns. They're the ones who treat the evidence and the deadlines with as much seriousness as actually spending the money itself.
Molly Ridley, Head of Partner Growth & Marketing, Noteworthy SupportGood news, really. It's more of an admin problem than a marketing one, and once it's set up it costs very little to maintain.
Questions from the session
We've previously funded trips to Ignite through co-op. Is that still allowed?
Yes. There are caps on train, hotel and airfare, but it's still claimable, and Ignite is specifically named in the guidelines as a standout event.
On paid digital media, where's budget best focused? We're currently on LinkedIn but open to suggestions.
It comes back to where your customers actually are. If you haven't tried a channel before and you think your audience might be on it, run a small test. You don't need to commit £10,000; a couple of hundred pounds tells you whether the impressions, the engagement rate and the targeting stack up. If LinkedIn's working for a specific customer set, say CTOs in 100-200 seat retail businesses, the closest thing for that level of targeting is probably Microsoft Ads. Microsoft owns LinkedIn, and it's a badly underused platform for partners. We've done a lot of trends analysis here and certain keywords and products have far higher search volumes and better engagement in Bing than Google. The display side lets you hone right in, so you know your £500 is only ever in front of those CTOs. Beyond that, we've had success with Reddit on specific search terms. ChatGPT paid ads are new, fairly expensive and quite broad, so they're hard to focus, but not a bad testing ground if your audience is wide. Don't be afraid to experiment. Just do it in a considered, low-risk way, and scale what works.
Missed it, or want to hear it in full?
The session ran just under an hour and there's plenty we couldn't fit into a writeup. Fill in the form and we'll take you straight to the recording.
- The worked Copilot campaign example
- The full funnel breakdown
- The live Q&A in full