Somewhere between your last board meeting and your last pipeline review, someone told you that you should be doing partnerships. So you opened a few guides. Plenty of them lead with infrastructure: choose a partner relationship management platform, define your partner tiers, build a portal, recruit.
That advice assumes a company you are not yet. If you are a SaaS founder with a few dozen customers, a pipeline that still leans on you, and nobody whose job this would be, a program is the most expensive way to find out whether partners will sell for you. This article argues for a smaller first step: one partner, one co-sold deal, about 90 days, and a decision rule written before you begin. Tooling, tiers and a partnerships hire come after one partner has put their name on one deal.
The standard advice assumes a team you do not have
Read what the vendors recommend and notice what they sell. Introw, a PRM vendor, publishes a guide to B2B SaaS partnerships that lists referral programs and integrations with other software vendors as good early plays, then spends most of its length on the platform capabilities a program needs to scale: tiering, commissions, portals, deal registration. That is a reasonable guide to running a program. It is not a guide to deciding whether you should have one.
The more useful warning comes from a practitioner quoted in PartnerStack's guide to co-selling. Nicolette Lopes, Head of Partnerships at Carbon6, describes companies that hand partnerships to someone already doing another job, "half your job now", and says partners can tell when a vendor is not investing in them. Note the source: PartnerStack also sells partner software. But the failure she describes is one most founders will recognize, because it is the default outcome when a program is started without an owner.
The data on partner involvement is encouraging and also messy. Crossbeam's 2024 analysis of its network found an average win-rate lift of 11.7% on deals where partners were involved, with a lift of 9.4% for companies with one to five connected partners. For smaller companies the path was uneven: the lift for SMBs fell from 8.7% at one to five partners to 4.6% at five to ten, before recovering in much larger ecosystems. Crossbeam's own explanation is a hypothesis, that partnership management hits "alignment and efficiency walls" before a program matures. Treat the numbers with care. They come from a vendor's network of self-selected users, count only partners connected on that platform, and show correlation rather than cause.
Put together, these sources suggest a practical reading. Partners can help a small company win, and adding partners faster than you can manage them is where the benefit leaks away. That is an interpretation, not a finding, but it points the same way as the operating advice: keep the number of relationships small enough to run properly.
You probably already have partners
The cheapest place to find your first partner is your own deal history. Lopes describes "shadow" co-selling: partners who already influence customers to buy from you, without anyone treating them as partners. Her examples are practical. Look in internal chat for threads where a customer wants to introduce you to someone. Look at implementations where the person doing the work has a different email domain from your customer, because that firm is helping your customers get value from your product.
Run that as an audit. Pull every closed-won deal and every serious stalled deal from the last twelve months. For each one, list every outside party who touched it: an implementation firm, an agency, an advisor, an integration vendor, a customer who made the introduction. In practice this tends to surface a short list of names that recur. That pattern is a reasonable expectation rather than a guarantee, so if the audit returns nothing, that is a finding too: you do not have a partner motion to formalize yet, and the next step is picking a partner type rather than a partner.
Rank what you find by how many deals each name touched and how many of those you won. We make the same argument about pricing: the segmentation data you need is already in the deals you have closed, stalled and lost. The deal history is a better prospecting list for partners than any directory.
Pick one partner and map the overlap by hand
Choose a single partner from the shortlist. Three tests decide it. Their customers should look like your ideal customer. Their work should sit next to yours rather than compete with it. And they should have a reason to care, which usually means you make their service easier to sell, easier to deliver or harder for a customer to leave.
Then map the overlap yourself. Put their named accounts next to your target list in a spreadsheet and mark where they meet. Keep the list short enough to walk through in one call with the partner. You do not need an account-mapping tool for this, and buying one before you have a partner who will join the call buys you software without a use for it.
The output of the call is a handful of accounts where both of you have a reason to act, and a name against each: who knows the buyer, and who should open the conversation.
Write the one-page deal plan
Co-selling means you stay in the deal and the partner adds credibility, context or technical authority. That is different from reselling, where the partner buys and resells your product and takes margin. The distinction matters because it decides who owns the client and who handles contracting. Settle it before the first conversation.
Add one more line: who on your side owns this, and how many hours a week they have. If the honest answer is "nobody", settle that first. A growth-partner engagement is one way to add capacity without a full-time hire, and the trade-offs of that choice are covered in our piece on the VP of Sales hire as a risk decision. If you prefer to run it yourself, put the hours in your calendar before you contact the partner.
Decide at day 90, using criteria you wrote on day one
Ninety days is a judgment, not a benchmark. It is long enough for one deal to move through a real conversation and short enough that you can stop without regret. What matters is that the decision rule exists in advance, because a partnership with no stop rule survives on goodwill.
Stopping is a good result. You have spent a few hours a week for a quarter and learned that this partner, for this buyer, will not act. You have not signed a platform contract, hired a manager or told other partners to expect a program.
What if one deal teaches you nothing?
The strongest objection to this approach is that partnerships need real investment, and one deal cannot validate a program. Lopes's warning applies directly: partners notice a half-committed vendor, and a burned partner rarely returns. Crossbeam's uneven results for small ecosystems suggest they carry costs of their own. A single deal, the objection runs, is too small to learn from.
Part of that is right. One co-sold deal does not prove you can run a program, and it should not be presented to a board that way. What it tests is narrower and comes first: whether a specific partner will attach their name to a specific deal for a specific buyer. If the answer is no, no portal will change it. If the answer is yes, you have a real case, a named owner and a partner who has seen how you sell, which is a much better starting point for a program than a signed agreement and an empty pipeline.
The approach does not apply if partners already source deals for you on a regular basis. At that point you are past the proof stage, and the question becomes how to run what you have.
Earn the program
Build the program when the results repeat. A second deal with the same partner, or a first deal with a second partner, is the signal, together with an owner who has real hours. The spreadsheet will tell you when it is time for a tool: when keeping it current costs more time than the partner conversations themselves, software pays for itself. A partnerships hire follows the same rule. Hire when the owner's hours have outgrown the role, not before.
Until then, the work is small and specific. This week, pull last year's closed-won deals and write down every outside name that touched one. If the list has a name on it more than once, you have found your first partner conversation.
