How do I decide between vertical-by-vertical vs horizontal expansion?

Direct Answer (TL;DR): Go vertical-by-vertical when (a) workflow logic, data model, or compliance materially differs by industry, (b) the buyer title only exists in that industry (Chief Nursing Officer, Director of Underwriting, Plant Manager), or (c) one segment has 3x the wedge TAM of the next. Go horizontal when the same UI, the same buyer title (CFO, CRO, Head of RevOps), and the same security review (SOC 2 Type II, GDPR) clear every industry you touch. Use the formal threshold from Bessemer's State of the Cloud reporting: net-new ARR per AE in segment 1 should reach the same payback as segment 0 inside 4 quarters; if it doesn't, you are bolting on, not expanding.

The Detail

Most "vertical vs horizontal" debates are mis-framed as a branding question. They are actually four separate questions stacked on top of each other: product, buyer, sales motion, and unit economics. You only have a real "vertical" if at least three of the four are different. Otherwise you are running a horizontal product with vertical marketing copy, which is fine, but should not be financed like a true vertical play.

Decision matrix (use all seven rows, not the top three):

If five or more rows fall on the Vertical column, run vertical-by-vertical. If five or more fall horizontal, run horizontal. A 4 to 3 split means hybrid; do not split the difference at 3 to 4 by accident.

Vertical-by-vertical playbook (the Veeva, Toast, Procore, ServiceTitan pattern):

1. Beachhead selection — Score every candidate vertical on (TAM_dollars * win_rate_estimate) / (CAC_payback_months * compliance_gate_months). Pick the highest. Veeva (founded 2007, IPO 2013 at ~$200M ARR) chose pharma life sciences because it had the largest dollar TAM with the longest compliance moat; Toast (S-1 filed 2021 at $704M ARR) picked full-service restaurants because card-present + tipping + back-of-house workflows were ignored by Square and Clover; ServiceTitan (S-1 2024) did the same in residential trades.
2. Vertical proof phase (months 0 to 12) — 2 to 4 AEs, target 6 to 12 logos. Build vertical-only assets: ICP firmographic list (industry NAICS code, employee count band, estimated revenue band), buyer persona deck, vertical case study, analyst inclusion (Gartner Magic Quadrant in the vertical sub-category, not the broad one), and a vertical-specific user community (Veeva Commercial Summit, Toast Spark).
3. Replication gate — Only replicate when (a) NRR >= 115% in the original vertical for 2 consecutive quarters, (b) at least 3 of the first 10 customers are referenceable with named ROI, and (c) win rate in qualified vertical pipeline is >= 30%. If any of the three fails, do not start vertical 2 — fix vertical 1.
4. Vertical 2 cadence — Hire 4 AEs, 1 vertical SE, 1 vertical CSM. Replicate the playbook in 9 months, not 18; if vertical 2 takes as long as vertical 1, the playbook is not transferable and you are running two separate companies.

Horizontal playbook (the Salesforce, Stripe, Datadog, HubSpot post-2018 pattern):

1. Problem-first, not industry-first positioning — Stripe never said "payments for SaaS"; it said "payments infrastructure for the internet." Datadog never said "monitoring for fintech"; it said "observability for any modern stack." The narrative anchor is the workload, not the customer's NAICS code.
2. Self-serve + product-led signal harvesting — A horizontal product can let the market self-segment. HubSpot's 2018 to 2022 expansion happened because freemium signups came from every industry; the GTM team simply followed the heat map. If you cannot get freemium or low-friction trial signal across industries, you may not actually be horizontal.
3. Pricing by company size, not industry — SMB / mid-market / enterprise tiers, with the same SKUs. Industry-priced horizontal products eventually get arbitraged by buyers who notice.
4. Specialist overlays — Once a horizontal motion is at $100M ARR, add 2 to 4 industry overlay SEs (FinServ, Healthcare, Public Sector) for deals over $250k. This buys vertical credibility without forking the GTM org.

Trigger conditions — which path is actually correct:

Go horizontal if:

• The same code, the same UI, the same security review serves any industry.
• The buyer title exists in 8+ industries (CRO, CFO, Head of Data, VP Engineering).
• Win rate variance across industries in your pipeline is under 10 percentage points.
• You already have product-market-fit signal in 2+ industries with under 15% feature differentiation between them.

Go vertical if:

• The data model, integrations, or compliance posture must materially change per industry.
• The buyer title only exists in one industry (Chief Nursing Officer, Director of Claims, Branch Operations Manager).
• One segment has 3x the wedge TAM and a longer compliance moat than the next.
• Your competitor in the segment is a horizontal generalist; vertical depth is the wedge.

Operating cadence (what the CRO actually runs every week):

• Vertical motion weekly review: pipeline coverage by named-vertical account list (not generic "healthcare"; the literal 200 accounts you've targeted), win rate by buying-committee composition, and stage conversion from SAL to Closed-Won. Coverage target 3.5x of segment quota.
• Horizontal motion weekly review: pipeline by company-size band (1 to 50 / 51 to 500 / 501 to 5000 / 5000+), self-serve to sales-assist conversion rate, and PQL (product-qualified-lead) to opportunity rate. Coverage target 3.0x.
• Exit condition for vertical 1: 3 consecutive quarters of NRR >= 115%, win rate >= 30%, and 5+ referenceable customers. Only then start vertical 2.
• Exit condition for horizontal: NRR drifts below 105% for 2 consecutive quarters, or new-logo win rate falls under 22% — at that point a vertical overlay is needed before more horizontal spend.

KPI dashboard (the 8 numbers the board should see monthly):

1. NRR by segment (vertical or company size)
2. Gross margin by segment
3. CAC payback by segment (months)
4. Win rate by segment, qualified pipeline only
5. Pipeline coverage by segment (next-quarter quota)
6. Time-to-first-value, days from contract to production usage
7. Logo retention, trailing 12 months
8. Average ACV by segment, with 90-day trend

Worked examples (ARR milestones to copy):

• Veeva CRM: 1 vertical (pharma) to $100M ARR before adding Veeva Vault in 2011; horizontal-within-vertical expansion, not cross-industry. Today >50% of revenue is Vault, but it took an 8-year vertical compounding base to fund it.
• Toast: 1 vertical (full-service restaurants) to $400M ARR before adding Toast Retail (2024) and Toast for Hotels. Same wedge replication, adjacent vertical only after the original was a fortress.
• Salesforce: horizontal CRM core, then vertical clouds (Health Cloud 2015, Financial Services Cloud 2015, Manufacturing Cloud 2019, Public Sector 2021). Each vertical cloud launched only after horizontal CRM was at $5B+ ARR.
• Stripe: horizontal payments primitives, vertical overlays (Stripe Connect for marketplaces, Stripe Climate, Stripe Atlas) layered on top of a single API surface. No vertical sales org until $7B+ ARR.
• HubSpot: horizontal SMB CRM, segmented by company size not industry. Industry-specific landing pages but a single GTM motion.
• Datadog: horizontal observability, with FinServ and Public Sector overlays added post-IPO at >$1B ARR.

Hybrid model (the most common honest answer):

Most successful expansion stories are hybrid in motion even if they brand as one or the other. Salesforce launched Health Cloud (2015), Financial Services Cloud (2015), and Public Sector Cloud (2021) on top of a horizontal core. Toast added Toast Retail (2024) on top of restaurant. The hybrid rule: own 1 to 2 verticals deeply with vertical AEs, vertical messaging, and vertical-specific compliance, then cross-sell horizontally to adjacent segments using shared platform leverage. Do not invest in horizontal expansion before the first vertical is at $25M ARR with 115%+ NRR; you will dilute the wedge.

Financial guardrails (do not skip these):

• Vertical play minimum gross margin: 70%. Below 70% you cannot fund the 14 to 18 month CAC payback at scale.
• Horizontal play minimum NRR: 105% gross retention floor, 110% net. Below 110% NRR a horizontal motion is a leaky bucket.
• Cash runway needed before launching vertical 2: 24 months at current burn, with vertical 1 fully self-funding (gross profit covers vertical 1 S&M + R&D allocation).
• Rule of 40 floor before splitting the GTM org: combined growth + FCF margin >= 30% pre-split, or you create two sub-scale orgs that both miss plan.

Bear Case (what goes wrong, and how often):

• Vertical playbook fails to replicate. Roughly 40% to 55% of "we'll do 5 verticals" plans stall at vertical 2 (per OpenView's 2023 expansion benchmarks) because the playbook was not actually documented; the founders were the playbook. Symptom: vertical 2 ramps as slowly as vertical 1, and AE attainment in vertical 2 sits below 60%.
• Horizontal pricing collapse. Horizontal products with no segment differentiation get commoditized; logo land-rate is high but NRR drops below 100%. If horizontal NRR is below 105% for two consecutive quarters, you are renting customers, not owning them.
• The "fake vertical" trap. A horizontal product with vertical marketing copy looks like a vertical play to the board, but win rate inside the "vertical" matches win rate outside it. Test: pull win rate in named-vertical accounts vs. non-vertical; if delta is under 5 points, you are horizontal with skins.
• Over-investment in compliance. FedRAMP Moderate authorization typically takes 12 to 18 months and roughly $2M to $3M of program cost (per GSA's FedRAMP PMO published guidance). Do not start FedRAMP unless you have 3+ named federal logos with signed letters of intent; otherwise the spend deadweights the P&L.
• Mis-timed split. Splitting the sales org into vertical pods before each pod has its own $10M ARR base creates sub-scale pods that all underperform a unified team. See q88 for the segmentation timing rule.

Related entries: /knowledge/q80 (rolling out a 15% price increase without churning the base), /knowledge/q85 (ICP segmentation for $10M ARR mid-market SaaS), /knowledge/q86 (SMB to mid-market expansion without breaking SMB), /knowledge/q88 (split sales org by segment vs region), /knowledge/q89 (trigger to launch an enterprise motion separate from mid-market), /knowledge/q90 (evaluate whether a new vertical is worth the GTM investment), /knowledge/q91 (realistic CAC payback for SMB vs mid-market vs enterprise), /knowledge/q92 (partner/channel motion alongside direct sales), /knowledge/q95 (federal / public-sector motion from scratch).

Common founder mistakes (in order of frequency observed):

1. Calling a horizontal product "vertical" because the marketing site has industry pages. The org chart, the AE comp plan, and the product roadmap all have to actually fork for it to be a vertical play.
2. Starting vertical 2 before vertical 1 is at NRR >= 115%. This is the single most common reason expansion plans miss plan by 30%+.
3. Hiring industry-credentialed AEs for a horizontal product. They sell slower than generalists because they over-customize the pitch; horizontal products win on speed and consistency.
4. Pricing horizontally but selling vertically. If your pricing page has one set of SKUs but every deal is custom, you have a horizontal product priced like SaaS but sold like services. Margins collapse.
5. Skipping the compliance gate analysis. FedRAMP, HITRUST, PCI Level 1, and FINRA each add 6 to 18 months of cycle time. A vertical play that ignores this misses Year 1 plan by 40%+.

Decision tree (90 seconds):

1. Does the product require an industry-specific data model or compliance audit? If yes -> vertical. If no -> step 2.
2. Does the buyer title only exist in one industry? If yes -> vertical. If no -> step 3.
3. Is win rate variance across industries in your existing pipeline > 15 points? If yes -> vertical. If no -> step 4.
4. Is one segment > 3x the TAM of the next? If yes -> vertical first, horizontal second. If no -> horizontal.

Self-check before adopting any of this:

• Do you actually have win-rate data segmented by industry, or are you guessing? If guessing, the first action is instrumenting the CRM, not picking a path.
• Have you separated logo CAC from expansion CAC by segment? Combined CAC hides which path is profitable.
• Have you reviewed compliance certificates per target industry with a security lead, not just sales? Sales will under-estimate audit cycle time.

Sources:

• Bessemer Venture Partners, State of the Cloud 2024 — https://www.bvp.com/atlas/state-of-the-cloud-2024
• OpenView 2023 SaaS Benchmarks Report — https://openviewpartners.com/2023-saas-benchmarks-report/
• Gartner Magic Quadrant for Vertical-Specific Cloud Suites — https://www.gartner.com/en/documents/4022991
• McKinsey, "The vertical SaaS playbook" — https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-rise-of-vertical-software
• ONC 2024 EHR Documentation Burden Report — https://www.healthit.gov/data/data-briefs
• GSA FedRAMP Program Office — https://www.fedramp.gov/program-basics/
• Veeva Systems S-1 (SEC EDGAR, 2013) — https://www.sec.gov/Archives/edgar/data/1393052/000119312513381390/d586505ds1.htm
• Toast S-1 (SEC EDGAR, 2021) — https://www.sec.gov/Archives/edgar/data/1650164/000162828021018466/toasts-1.htm
• ServiceTitan S-1 (SEC EDGAR, 2024) — https://www.sec.gov/Archives/edgar/data/1689948/000119312524262920/d745218ds1.htm

TAGS: market-expansion,vertical-strategy,horizontal-expansion,gtm-design,icp-segmentation