The growth question, finally
Every session so far earned the right to ask this one. We know why Himalaya stalled (activation, not market), what the outside pressure is (AI-natives onboarding in days), and what the redesigned offer looks like (a usage-priced starter tier under test). Session 2 parked the AI copilot as "Horizon 2" - today is where that placement gets its logic. Three lenses, one portfolio: what stage of life is this company in, which bets belong in which horizon, and which direction each move takes on the map.
The business lifecycle - and why stages break 7 min live
Companies move startup → growth → maturity → renewal or decline, and each stage has a characteristic way of dying. Larry Greiner's growth model (HBR, 1972) sharpened it: growth is not a smooth curve, it is a series of phases that each END IN A CRISIS. The fix that powers one phase manufactures the next crisis. Himalaya's weeks-long onboarding is not laziness - it is a lifecycle symptom.
LiveDiagnose your stage honestly4 min▶
Stage diagnosis is the anchor because the fatal error in growth planning is running the wrong stage's playbook - a maturity playbook in growth (optimize, cut, standardize) or a growth playbook in maturity (spend, hire, sprawl). Three cheap instruments read the stage:
- Revenue growth rate and its trend. Himalaya: 45% → 12% in three years. Not decline - decelerating growth. Classic late-growth signature.
- Decision latency. How long from question to committed answer? Session 2 opened with a three-way deadlock between product, sales, and CS - deadlock is the control crisis knocking.
- Founder / exec proximity to customers. When leadership last heard a live customer complaint firsthand is a surprisingly precise stage clock.
The playbook mismatch that cost a year: a scale-up brought in a big-company COO who installed quarterly planning gates, approval chains, and a PMO - maturity medicine for a growth-stage patient. Velocity halved, the best builders left, and the board spent the next year undoing it. The tools were fine; the stage diagnosis was wrong.
LiveWhat each stage rewards3 min▶
Each stage pays out on a different virtue - and quietly punishes last stage's virtue:
| Stage | Rewards | Punishes |
|---|---|---|
| Startup | Learning speed - cheap experiments, fast kills | Premature process |
| Growth | Repeatability - a sales motion and onboarding you can clone | Hero-dependence |
| Maturity | Efficiency + renewal bets funded from the core | Growth-at-all-costs spending |
| Decline | Honesty - harvest, pivot, or exit deliberately | Nostalgia budgeting |
Himalaya's growth stage rewards repeatability - which is exactly what the partner-reseller onboarding gap (Session 3's root cause) broke. The activation playbook is stage-correct medicine.
Self-studyS-curves: every engine flattens3 min read▶
Richard Foster's S-curve is the lifecycle's physics: any technology or business model yields slow gains early, steep gains in the middle, and diminishing returns at the top - no matter how well you execute. The strategic skill is jumping to the next curve BEFORE the current one flattens, while the core still funds the jump.
- The trap: at the flat top, more effort produces less result, so teams respond with... more effort. Deceleration despite good execution is the tell - and 45% → 12% with a stable team is textbook.
- Himalaya's read: the platform S-curve is flattening exactly as the AI curve steepens - the two AI-native competitors are not better at Himalaya's curve, they started on the next one. That is why the copilot bet exists at all.
- Pairing: Everett Rogers' adoption curve tells you WHO arrives at each point of the S - early adopters forgive rough edges, the mainstream will not.
Three Horizons: hold three clocks at once 5 min live
McKinsey's Alchemy of Growth answer to the flattening curve: run three timeframes simultaneously. Horizon 1 defends and extends the core (12-18 months), Horizon 2 scales emerging bets (2-4 years), Horizon 3 keeps cheap options on the future (4+ years). Growth dies when a company runs only one clock.
LiveBudget and metrics per horizon - the actual discipline3 min▶
Three Horizons fails as a poster and works as a governance rule: each horizon gets its own budget line AND its own yardstick.
- H1 is judged on margin. The activation fix and starter tier must show contribution and NRR movement inside 12-18 months. No poetry allowed.
- H2 is judged on trajectory. The AI copilot is measured on adoption slope, pilot conversion, usage depth - is the curve bending up? - not on this year's P&L.
- H3 is judged on learning. The consumer-side option buys answers to named questions (will consumers engage directly? is the data loop real?) at a deliberately small cost.
- The classic failure: judging H3 options by H1 math. Ask a 4-year option for a 12-month ROI and you will kill it every time - then wonder, four years later, why there is nothing in the pipeline.
The annual budget massacre: a platform company ran every initiative through one ROI template each November. For five straight years, everything beyond an 18-month payback scored last and got cut - rational each time. Year six, the core flattened and the shelf of ready bets was empty. Nobody decided to stop investing in the future; the single yardstick decided it for them.
Live70 / 20 / 10 - a starting argument, not a law2 min▶
A rough allocation heuristic: ~70% of growth resources to H1, ~20% to H2, ~10% to H3. Its value is not precision - it is that it forces the allocation conversation to happen at all, with explicit numbers someone can argue against.
- Tilt by stage: a flattening core (Himalaya) argues for nudging H2 up; a bleeding core argues for overweighting H1 until the leak is fixed - which is exactly Himalaya's first move.
- The test of seriousness: if H2 and H3 have no named budget and no named owner, you do not have horizons - you have a slide.
Self-studyBCG growth-share matrix: when you have MANY products3 min read▶
Bruce Henderson's 1970 matrix plots each product on market growth × relative market share: stars (high/high - invest), cash cows (low growth, high share - milk to fund the rest), question marks (high growth, low share - pick a few, kill the rest), dogs (low/low - divest or ignore). It is a portfolio lens for when Three Horizons' three buckets are not granular enough.
- Himalaya's modules as a mini-portfolio: payments and storefront behave like cash cows (high attach, low growth), the engagement module like a question mark (low attach, but multi-module clients that have it retain best), the copilot as a future star candidate.
- Handle with care: the matrix assumes share drives cost advantage (experience curve) - true in 1970s manufacturing, only sometimes true in software. Use it to structure the funding debate, not to settle it.
Ansoff: the direction of growth 5 min live
Igor Ansoff's 1957 matrix asks the simplest growth question: existing or new products, sold to existing or new markets? Four directions, rising risk as you leave home. It is the map that turns Himalaya's pile of initiatives into a sequence.
LiveRisk rises as you leave home - so sequence3 min▶
- Penetration is cheapest because everything is known: same product, same buyers, same channel. Himalaya's activation fix and starter tier live here - the failure modes are already diagnosed (Session 3) and the offer is already designed (Session 5).
- One new variable at a time. Product development (copilot) changes the product but keeps 1,200 known clients as the launch market. Market development changes the buyer but keeps the product. Diversification changes both - which is why it enters the portfolio as an H3 option, not a commitment.
- The funding chain: penetration recovers NRR → recovered NRR funds the copilot → a working copilot deepens the consumer data loop that makes diversification even thinkable. Break the chain's first link and everything downstream starves.
- Parked is a position too: new verticals/geographies stay parked deliberately - Ilsa's enterprise push lost the Session 2 matrix, and Ansoff explains why: it was a new-market move attempted while the home quadrant was still leaking.
The double-jump: a mid-market SaaS firm, core flattening, leapt straight to diversification - a new product for a new buyer, sold as "the bold move." Two unknowns compounded: they mispriced the product AND misread the channel, and there was no home-quadrant cash engine left to cover the tuition. The bold move was actually two moves, taken at once, on borrowed money.
Self-studyTAM / SAM / SOM: size the starter tier honestly3 min read▶
Before the starter tier gets real budget, size it: TAM (everyone who could conceivably buy), SAM (the slice your product and channel can actually serve), SOM (what you can realistically win in 2-3 years). The discipline is in the method, not the acronym.
- Top-down lies: "the SMB software market is $80B, 1% of that is $800M" is a spreadsheet hallucination - no mechanism connects you to the 1%.
- Bottom-up counts buyers: how many businesses fit the starter-tier profile, in reachable channels, at what conversion, paying what? Numbers a chairman can interrogate line by line.
- Himalaya sanity check: 1,200 clients came from a definable universe of mid-size consumer businesses; the starter tier widens that universe downmarket. Count the widened ring, not the whole market.
Self-studyThe flywheel: growth as a loop, not a ladder3 min read▶
Jim Collins' flywheel (made famous by Amazon's napkin sketch) reframes growth as a self-reinforcing loop: each turn makes the next turn easier. Lifecycles and horizons say WHEN; the flywheel says what compounds.
- Himalaya's candidate loop: more consumers → more behavioral data → better platform recommendations → better client results → more clients → more consumers. If real, this is the moat the AI-natives cannot copy by being cheap.
- The H3 connection: the consumer-side option is really a test of whether this flywheel spins - which is why it is judged on learning, not revenue.
- The test: a real flywheel has no "and then marketing pushes harder" step. If a link needs constant manual shoving, it is a funnel wearing a flywheel costume - Session 7 takes that distinction further.
Place Himalaya on the curve ★ 8 min · every table diagnoses
Whiteboard: sketch the lifecycle curve with Greiner's four crisis bolts. Two minutes as a table: mark where Himalaya sits, with evidence from the brief - not vibes. Growth 45% → 12%, NRR 98%, churn concentrated, win rate sliding.
Name the crisis symptom you see. Candidates from the case so far: onboarding takes weeks (process drag), partner enablement descoped under growth pressure (Session 3's root cause), the three-way decision deadlock that opened Session 2 (control crisis knocking).
Each table states its diagnosis in one sentence: "Himalaya is at LATE GROWTH, entering the control / red-tape zone, and the evidence is ___."
Stress test: what would you expect to see if Himalaya were actually in DECLINE instead? (Shrinking ARR, multi-module churn too, logo losses across cohorts.) Confirm the brief says otherwise - the stage diagnosis holds.
Build the three-horizon portfolio ★ 8 min · every table allocates
Three columns on the whiteboard: H1 / H2 / H3. Place Himalaya's moves: activation fix + starter tier → H1, AI copilot → H2, consumer-side network effect → H3. Fast - the placements are Session 2 and 5 canon. The argument is what comes next.
Argue the split: start from 70/20/10 and let each table defend a deviation. Flattening core says push H2 up; leaking core says overweight H1 first. There is no single right split - there IS an indefensible one (100/0/0).
Assign the yardstick per horizon, written under each column: H1 = margin and NRR recovery, H2 = adoption trajectory, H3 = named learning questions answered per dollar.
The killer check: pick any H3 idea and ask "would this survive our standard ROI template?" If no - and it should be no - write down what WOULD kill it legitimately (the learning questions coming back negative).
The moment this lands with boards is the yardstick row, not the allocation row. "We will judge the copilot on trajectory and the consumer option on learning, and here are the kill criteria" sounds like governance. "Trust us, it is strategic" sounds like a wish.
Map every initiative onto Ansoff ★ 6 min · every table sequences
Draw the 2x2. Place everything Sessions 2-5 produced: activation fix, starter tier, AI copilot, enterprise push (deferred), consumer-side monetization. One sticky note per initiative.
Sanity-check the sequence against the risk gradient: penetration first (activation + starter tier), product development second (copilot), diversification held as an option. Does each move fund the next? Say the chain out loud.
Notice the coherence: the Ansoff sequence, the horizon portfolio, and the lifecycle diagnosis all point at the same first move. When three independent lenses agree, you have a plan instead of a poster.
This week ◐ 30 min total
- Place YOUR company on the lifecycle curve - growth rate trend, decision latency, exec proximity to customers - and name the coming crisis in one written sentence. Ten minutes, brutally honest.
- Three-horizon your own portfolio: list live initiatives, sort into H1/H2/H3, write the yardstick per horizon. Note where H2/H3 have no owner or budget - that gap IS the finding.
- Read the TAM/SAM/SOM card above, then bottom-up size one growth idea you currently believe in. Compare against the number you had been repeating.
- Optional deep end: Greiner's "Evolution and Revolution as Organizations Grow" (HBR 1972, short) or Baghai, Coley & White's The Alchemy of Growth chapters 1-2.
Three questions before you go 🎯 ◐ 90 seconds
1 · Himalaya's onboarding takes weeks, and every fix adds another checklist. Greiner would call this...
Each Greiner phase's solution manufactures the next crisis. Processes added to scale the growth phase now slow it - the answer is evolving the operating model, not blaming the people running the old one.
2 · The board asks the H3 consumer-side option for a 12-month ROI forecast. The disciplined response is...
Different horizons, different yardsticks: H1 margin, H2 trajectory, H3 learning. The classic failure is one ROI template for all three - it always allocates 100% to the present.
3 · Rank Himalaya's moves from LOWEST to highest risk on Ansoff:
Risk rises as you leave the home quadrant. Penetration changes nothing but execution; product development changes one variable (the product); diversification changes two (product AND market) - which is why it stays an option until the earlier moves fund it.
Frameworks covered & their origins
This course teaches the working 80% of each framework and cites the canon - the original books and papers stay the deep end for anyone who wants it.