How a Tiny Team Ships Like a Big One · The Partner Track
Co-sell is a data problem before it's a sales problem
A customer's co-sell program stalled for a year. The fix wasn't sales motivation - it was that nobody could answer 'which of our deals have AWS in them?' from the CRM.
One of our customers ran a co-sell program for about a year before we met them, and it produced single-digit ACE submissions. The postmortem inside the company blamed the usual suspects: reps don't see the value, alliance team has no leverage, AWS is hard to work with. Then their alliance lead asked us a question during onboarding that explained the whole year: "Can your product tell us which of our deals have AWS in them?" Nobody could answer it from the CRM. Not approximately, not at all.
Think about what co-sell actually requires you to know, per deal: is the customer on AWS, what's the workload, who on the customer side owns cloud decisions, and what stage is the deal really at. None of those are fields reps fill reliably, and two of them (the workload and the cloud stakeholders) usually live in email threads and call transcripts, not in the CRM at all. We measured that gap directly at another customer: 0.74% of emails attached to the opportunity they were about. An alliance lead trying to build a co-sell pipeline on top of that isn't running a sales program. They're running an archaeology program with a quota.
The order of operations
So the co-sell fixes that work run in an unglamorous order. First the deal record has to become true: communications resolved to deals, stakeholders mapped, the AWS-shaped facts (platform, workload, cloud owner) extracted from where they actually live. This is the same entity-resolution machinery the product track keeps writing about; co-sell turned out to be one more consumer of it. Second, eligibility becomes something computed, not remembered: every deal carries its own answer to "should AWS know about this?" Third, and only third, the sales motion: the one-click submission, the AWS intro, the joint call.
Run in that order at the customer with the stalled program, the pipeline that had produced single-digit submissions in a year turned out to contain 39 co-sell-eligible opportunities worth roughly $3.2M in potential AWS funding. The reps hadn't been unmotivated. The deals had been invisible.
The inverse order is the common one, and it fails predictably: alliance teams buy co-sell training, mandate ACE quotas, and put "AWS fit" as a required CRM field, all of which press harder on data that isn't there. The reps comply the way reps comply with required fields, which makes the data worse, which makes the program's numbers less trustworthy, which is roughly where that year-long postmortem started.
Real numbers
- Single-digit ACE submissions from one customer's first year of co-sell effort; 39 eligible opportunities ≈ $3.2M in that same pipeline once eligibility was computed.
- 0.74% email-to-deal association - the substrate a typical co-sell program is built on.
- 4 facts per deal that co-sell actually needs (platform, workload, cloud owner, real stage); 0 of them reliably in CRM fields.
Where the humans sit
Data can make a deal visible; it can't make a rep care. What changed the reps' behavior at that customer wasn't the dashboard, it was the first deal that came back from AWS with an account team attached and funding behind it. Humans respond to the first proof, and the alliance lead's real job is manufacturing that first proof fast. Everything the machines do here is in service of shortening the distance to it.
Steal this
Before spending anything on co-sell enablement, take your last quarter's closed-won deals and hand-answer the four facts for each: platform, workload, cloud owner, stage. If you can't do it from the CRM in an afternoon, you have a data problem, and no amount of ACE training will out-sell it. Fix the resolution first (build or buy), compute eligibility across the whole pipeline rather than nominating deals by memory, and spend the enablement budget celebrating the first funded deal loudly instead.
Next on this track: Marketplace, starting with the feature we built - listing and private offers as product workflow, not a one-time project.
This post is part of the partner track of How a Tiny Team Ships Like a Big One, a series on how six builders run a production AI company. Building at Aithon - if this is how you want to work, talk to us.