Project — Org Design for a 20+ Person Product Org¶
Meridian's product org has crossed 22 people (18 PMs plus Amara and the four other Leads counted separately from their pods, plus Kai from the RouteWise acquisition now folded in as a Senior PM under Ravi). CEO Lena asks Amara for a full org design document ahead of next year's planning — not a status update, a real structural proposal she can take to the board alongside the budget ask. This project is the full deliverable: the kind of document a VP Product actually produces once a year, built from every tool in this level.
1. Current-state map¶
| Pod | Lead | PMs | Product line(s) | Posture (Module 2) |
|---|---|---|---|---|
| Onboarding & Journey | Ravi | 5 (incl. Kai) | Core Carrier Ops | Core |
| Pricing | (vacant — Sofia covering) | 3 | Core Carrier Ops | Core |
| Integrations/Platform | Tom | 3 | Shared infrastructure | Platform (Module 8) |
| Marketplace | Sofia | 4 | Marketplace | Growth |
| Internal Tools | Priya | 3 | Internal Tools | Seed/Harvest |
That's 18 PMs across 5 Leads (one Lead role currently double-covered by Sofia), consistent with Module 3's span-of-control numbers.
2. Diagnosis: three structural problems¶
| Problem | Root cause (Module 3 diagnosis) | Evidence |
|---|---|---|
| Sofia is overloaded — Lead of Marketplace and covering Pricing | Growth-bet Lead absorbed a vacant Core role without capacity planning | Sofia's 1:1s with her own PMs have slipped from weekly to biweekly per her own report |
| Tom's Integrations pod is structurally a platform team wearing a product-team org shape | Never formally re-typed after organic growth into shared infrastructure | 80% of Tom's backlog is cross-pod requests (Module 8) |
| Onboarding & Journey pod is oversized relative to peers | Kai's RouteWise team folded in without a re-split | 5 PMs vs. peer pods' 3-4, pushing average span toward the ceiling (Module 3) |
3. Target structure¶
python3 -c "
current = {'Ravi': 5, 'Pricing(vacant)': 3, 'Tom': 3, 'Sofia': 4, 'Priya': 3}
total = sum(current.values())
print('current total PMs:', total, 'across', len(current), 'pods')
# Target: split Ravi's oversized pod, hire dedicated Pricing Lead,
# formalize Tom as Platform with its own tax-funded headcount (Module 8)
target_pods = {
'Onboarding': 3, # Ravi keeps onboarding core
'Journey/RouteWise': 2, # Kai promoted to lead this narrower scope, per M&A leveling plan
'Pricing': 3, # new dedicated Lead hired
'Platform (Tom)': 3, # unchanged headcount, re-typed funding model
'Marketplace': 4, # Sofia back to single-pod focus
'Internal Tools': 3,
}
print('target total PMs:', sum(target_pods.values()), 'across', len(target_pods), 'pods')
print('avg span:', round(sum(target_pods.values())/len(target_pods), 2))
"
Result: 18 PMs currently across 5 pods (avg span 3.6) become 18 PMs across 6 pods (avg span 3.0) — the same headcount, redistributed, plus one new Lead hire (Pricing) and one internal promotion (Kai, per the M&A leveling plan from Module 7).
4. The three decisions this proposal asks the board/CEO to make¶
- Approve one external Lead hire (Pricing) — cost: one Product Lead salary, no net PM headcount change since the pod is a split of existing scope, not new scope.
- Approve Kai's promotion to Product Lead of the RouteWise-derived Journey pod, contingent on the two-quarter evidence review from Module 7's leveling plan completing on schedule.
- Approve re-typing Tom's pod as Platform with a 20% engineering-tax funding model (Module 8), removing it from portfolio competition against revenue-generating lines.
5. Risk and rollback plan¶
| Risk | Mitigation | Rollback trigger |
|---|---|---|
| New Pricing Lead hire takes >1 quarter to fill | Sofia continues covering with an explicit end-date and reduced Marketplace targets during the gap | If unfilled after 2 quarters, reassess whether Pricing needs a dedicated Lead at all |
| Kai's promotion evidence doesn't clear the bar in 2 quarters | Kai remains Senior PM under Ravi, Journey pod folds back into Onboarding | Explicit review date already set (Module 7) |
| Platform tax model creates funding resentment from product Leads | Amara reviews tax rate quarterly against actual platform adoption metrics (Module 8) | Adoption metrics flat or declining for 2 straight quarters |
6. Communication plan¶
Following Module 5's structure, Amara pre-wires (Module 9) the two board members most likely to have questions — the lead investor's rep (unit economics of the new hire) and the founder-friendly director (whether this dilutes focus) — before the planning meeting, then presents the proposal to the full board using headline/ask/evidence/risk, not a walkthrough of this full document (which goes out as the pre-read).
How It Actually Works¶
The reason this proposal has to name specific decisions for the board or CEO to approve, rather than presenting a structure as a fait accompli, is that org redesigns fail most often not from a flawed structure but from contested legitimacy — a structure imposed without visible executive buy-in gets quietly undermined by whichever manager lost authority in the reshuffle, who routes around the new lines by escalating to their old relationships instead of the new formal owner. Naming the specific decisions being asked for forces the sponsoring executive to actually commit their authority publicly behind the new lines, which is what gives a reorg teeth in its first uncomfortable test — when someone tries to escalate around the new structure in month two, as they always do, the executive's prior public commitment is what makes them decline the end-run rather than quietly relitigate the whole design. Rollback plans matter for a parallel reason: naming in advance what evidence would trigger a reversal makes the redesign falsifiable and therefore credible, whereas a redesign with no defined failure condition tends to persist purely through sunk-cost and political inertia even after it's visibly not working.
Stretch goals¶
- Build the actual career-ladder leveling packet (Module 4 format) you would use to evaluate Kai's promotion at the two-quarter review point, with named evidence criteria decided now, before the outcome is known.
- Draft the board pre-read slide (Module 5/9 format) for this proposal, under 200 words, and have someone unfamiliar with Meridian read only the headline and ask — check whether they can correctly state what's being requested.
- Extend the target-structure headcount model to simulate what happens if Marketplace hits its growth targets a year early and needs to split into two pods — re-run the span-of-control math and identify which existing Lead is the natural candidate to lead the split-off pod.