Co Selling Strategy

Co-Selling Strategy

Define how direct sales and partner sellers collaborate on shared deals: the policy layer, never the deal-by-deal execution. This skill designs three artifacts as one coherent operating charter, and validates it with the user section by section:

• Deal registration policy - how a partner formally claims an opportunity and what that claim grants.
• Credit and attribution model - who gets counted and paid when a deal has more than one parent.
• Rules of engagement - between direct and partner sellers.

Co-selling is a B2B discipline with no true B2C analogue. It presupposes named accounts, quota-carrying reps, a consultative sales cycle, and an opportunity record - none of which exist in a consumer transaction.

The consumer-side counterparts are trade marketing, co-op advertising funds, and shelf-placement/channel agreements: a different discipline with different mechanics, so never force-fit deal registration or comp neutrality onto retail partnerships. The one part that transfers to any context, B2B or B2C, is the governance discipline itself: unambiguous ownership rules, documented tiebreakers, published criteria.

Interview

Ask before proposing anything. One question per message; offer multiple-choice options when possible. Skip anything the user already answered. If your harness has persistent memory, check it for a previously approved charter first and confirm what changed instead of re-interviewing.

• What does the direct sales org look like - segments covered (SMB / mid-market / enterprise), inside vs field, how quota is structured?
• Which partner types are in play or planned: referral, reseller/VAR, distributor, SI/consultancy, ISV/tech, marketplace/hyperscaler?
• What are ACV, average sales-cycle length, and direct win rate? (Cycle length sets the registration protection window.)
• When did an AE last close a co-sell deal and call it a good experience - and what made it good?
• Does the current comp plan create any friction against attaching a partner to a deal?
• How does an AE find out which partner to call for a specific account today, and how long does that take?
• If an AE decided to ignore the partner motion entirely, what would happen?
• What is the goal: partner-sourced pipeline, influenced revenue, marketplace transaction volume, reaching segments direct can't?
• Is there a working attribution record today - timestamped registrations, or credit claims made at close?

Ask these three last, always. Every ranked menu later in this skill is re-ordered against the answers, and by the time the reader reaches a menu they have already committed to a path.

• By what date must this motion show a result, and what has to be true by then? A date inside one quarter deletes the heavyweight JBP track and promotes the one-hour registration and lock-window fixes. A date beyond two quarters keeps comp-plan work in play.
• One-off unblock, or a compounding asset? One-off promotes registration policy and rules of engagement, which pay out on the next deal. Compounding promotes comp neutrality and attribution co-ownership, which pay back every quarter and almost nothing in the first one.
• What is the effort ceiling - hours available, whether sales leadership will reopen the comp plan at all, and how reversible the change has to be? A refusal to reopen comp deletes every comp funding rung outright. A demand for reversibility demotes anything published to partners, since a granted discount is far harder to withdraw than an internal rule.

Workflow

1. Run the Interview. If you can browse the web and the user has a published partner or deal-registration page, read it - audit reality before designing on a blank page.
2. Gate on readiness before designing anything. Decline, with reasons, when any disqualifier holds:

• The ICP can't be clearly articulated yet.
• Direct sales hasn't proven it closes reliably on its own.
• Positioning, pricing, or process is still shifting.
• ACV sits below roughly $3-5k, the practitioner floor for any sales-assisted motion (general SaaS consensus, not a co-sell-specific study).
• Partnerships are being reached for as a fix for a growth problem.

Partners amplify a working direct motion. They do not fix a broken one. When the gate fails, recommend two or three deeply engaged referral partners with no formal program, and stop.

See references/partner-types-and-process-tracks.md for deeper diagnostic questions for this gate.

1. Pick the process track. Present both in ranked order with the trade-off in each, and recommend explicitly. The axes disagree:

• efficiency: lightweight playbook > heavyweight JBP
• value: heavyweight JBP > lightweight playbook
• effort: heavyweight JBP > lightweight playbook

Lightweight playbook fill: one fixed template filled per partner in a single 90-minute session. Buys a named owner per cell and a shared deal-stage map, the two things most disputes actually turn on, for one session per partner plus a one-time template. Best ratio, and the default for every partner until something moves it.

Heavyweight joint business planning: exec kickoff on both sides, shared measurable objectives, multi-week drafting, formal sign-off, quarterly reviews. Buys two-sided executive sponsorship and a budgeted plan, the only lever that moves a partner whose own field sellers carry quota, for weeks of drafting, a two-sided sign-off chain, and a standing quarterly review job. Highest value, worst ratio - move up to it only for a partner whose sellers you need to influence directly and where both sides will fund the review cadence.

Delete the JBP track from the menu, rather than recommending against it, when the partner side will not name an executive sponsor: an unsponsored joint plan produces a document nobody executes. Run lightweight and revisit when a sponsor exists.

Read references/partner-types-and-process-tracks.md before this step - the ordering above is a default that shifts with partner type, ACV, and who on the user's side actually runs the session.

1. Map each partner type in play to the registration and credit mechanic its economics actually require (same reference):

• A referral partner needs clean attribution, not margin protection.
• A reseller needs formal registration.
• An SI is motivated by services pipeline.
• An ISV/tech partner needs none of the resale mechanics.

Never ship one undifferentiated policy across types.

1. Design the five policy decisions in order, one at a time. For each decision:

• Present 2-3 candidate approaches ranked by value per unit of effort, never cheapest-first.
• Fold each option's trade-off into its own line and state an explicit recommendation.
• Get the user's call before moving on.
• Say the ordering out loud with a > line, and emit a separate line per axis wherever cost, value, effort and efficiency disagree.
• Express effort as hours, meetings, sign-off chains and reversibility, never as a currency amount.
• Drop any option the user's answers to the last three interview questions rule out, instead of demoting it, and say which answer deleted it.

Every ordering here is a default: re-rank it against what you already know about this user before presenting it.

1. Attribution taxonomy - partner-sourced vs partner-influenced, and whether to also publish a combined "partner-attached" number. See references/attribution-and-compensation.md.
2. Deal registration policy - what a registration grants, protection window and extension rules, approval criteria and SLA, published rejection grounds. See references/deal-registration-policy.md.