top of page
Won by pricing conversion-rate testing as a marketing subscription instead of a developer tool, so a CMO could approve it without an engineering budget request.
1
MODEL
BUSINESS MODEL
SaaS
model bm
HOW THEY BUILT IT
- Founded 2009 in Paris as Liwio, a web-analytics consulting agency; pivoted to a self-serve SaaS product in 2012 once the founders saw the same testing workflow recurring across client engagements.
- Bootstrapped until 2014, then raised Series A ($6M, 2016), Series B ($17M, 2018), and Series C ($40M, 2020) - total $64M - funding a new international office with each round.
- Now serves 1,100+ brands including Kering, L'Oreal, McDonald's, and Disneyland Paris.
HOW TO ARCHITECT IT
1. Run the service manually first (as a consultancy) because it teaches you exactly which steps to automate before you write a line of product code.
2. Price against the business outcome (conversion lift), not the technical capability (A/B testing), so budget owners outside engineering can say yes.
3. Land with one narrow, low-risk use case (website testing) before expanding into adjacent capability (personalization, feature flagging) - a contained pilot is easier for procurement to approve than a platform pitch.
4. Open a physical office in each new geography as you fund it (UK/Germany/Spain post-Series A, US/Singapore post-Series B) rather than trying to sell cross-border remotely from day one.
DISTRIBUTION MODEL
Enterprise Sales, Direct Sales
dm
HOW THEY OPERATIONALIZED
- Direct sales team targets marketing and product leaders at mid-to-large consumer brands, not developers.
- International expansion followed funding rounds: UK/Germany/Spain offices opened after the 2016 Series A, US/Singapore after the 2018 Series B.
- More recently added a Shopify App Store listing, layering a self-serve SMB channel on top of the enterprise sales motion.
HOW TO REPLICATE WHAT WORKED
What worked: converting a services practice into a repeatable SaaS product only after the workflow had proven itself across real client engagements, then using each funding round to open one new geography rather than over-expanding on the balance sheet.
The trap: don't mistake 1,100 enterprise logos for product-led growth. AB Tasty's engine is sales-cycle dependent, which is cash- and headcount-intensive. A founder copying 'get big logos' without also building the direct-sales team to close and service them will stall.
| PATTERNS OF THIS MODEL
PATTERNS IN CONSULTANCY-TO-PRODUCT TRANSITIONS:
1. RUN THE SERVICE MANUALLY FIRST. Delivering it by hand teaches exactly which steps to automate before a line of product code is written.
2. PRICE AGAINST THE BUSINESS OUTCOME, NOT THE TECHNICAL CAPABILITY, so budget owners outside engineering can approve it.
3. LAND WITH ONE NARROW, LOW-RISK USE CASE BEFORE EXPANDING INTO ADJACENT CAPABILITY. A contained pilot passes procurement; a platform pitch does not.
4. FUND EACH GEOGRAPHY WITH ITS OWN ROUND AND ITS OWN LOCAL PRESENCE. Selling cross-border remotely underperforms in categories where buyers expect local accountability.
What companies with this model reveal
| OPPORTUNITY INTELLIGENCE
GOLDMINE 1 — RUN THE SERVICE MANUALLY BEFORE YOU PRODUCTISE IT.
Standard: founded 2009 as a web-analytics consultancy, AB Tasty pivoted to SaaS in 2012 only after seeing the same testing workflow recur across client engagements. Consulting first tells you exactly which steps to automate before you write product code.
GOLDMINE 2 — PRICE AGAINST THE BUSINESS OUTCOME, NOT THE TECHNICAL CAPABILITY.
Standard: selling conversion lift rather than A/B testing puts the budget with a business owner outside engineering, where approval is faster and the number is bigger.
GOLDMINE 3 — LAND WITH ONE CONTAINED USE CASE.
Standard: website testing is easier for procurement to approve than a platform pitch; personalisation and feature flagging expand later inside the same account.
THE PIT — OPENING A PHYSICAL OFFICE PER GEOGRAPHY IS EXPENSIVE AND HARD TO REVERSE.
Each of the Series A, B and C funded a new market. That works while growth holds and becomes a fixed cost base in a downturn, in a category where Optimizely and VWO compete globally without the same footprint.
THE SECOND PIT — EXPERIMENTATION BUDGETS ARE DISCRETIONARY AND CUT EARLY.
MOVE WITH CAUTION — STATSIG AND WAREHOUSE-NATIVE RIVALS ARE REPRICING THE CATEGORY ON A UNIFIED DATA MODEL.
Untapped Business Model / Gaps / Goldmines / Pits
Patterns & Insights
2
MARKET
mkt mt es
MARKET TYPE
Fragmented Market
WHY THEY WON
No single vendor owns enterprise conversion-rate optimization - Adobe Target, Optimizely, VWO, and Dynamic Yield all compete for the same budget line. AB Tasty won share not by out-featuring Adobe's broader Experience Cloud but by being faster to deploy and easier to buy for a mid-market marketing team. The transferable principle: in a fragmented category, you rarely need to beat the biggest incumbent on features - you need to be the obvious 'good enough, faster to buy' option for the segment the incumbent under-serves.
ENTRY STRATEGY
Greenfield Entry
EXECUTION
AB Tasty had no outside funding until 2014, five years after founding, so its 2012 SaaS pivot was self-financed and untested by any partner or channel - a direct entry into the French e-commerce and retail segment using its own consulting client base as the first customers, evidence being that Series A only arrived once that direct-sold base was already proven.
FOOTHOLD STRATEGY
fs
Beachhead Strategy
French e-commerce and retail brands were the initial beachhead, drawn from the founders' existing consulting relationships. Once product-market fit repeated across that group, the company expanded geography by geography - UK, Germany, and Spain next, then the US and Singapore - rather than jumping straight to a global launch. Each new country was entered only once the prior one produced a repeatable playbook.
GROWTH CAMPAIGN
CAMPAIGNS THAT WORKED
- Customer success stories and case studies built around named enterprise logos to establish credibility with risk-averse buyers.
- Regular product webinars and live demos to educate marketing and product teams on the platform's capabilities.
- SEO-optimized content marketing (blog posts, whitepapers, e-books) aimed at digital marketers and product teams searching for conversion-optimization solutions.
KEY LEARNING
If you're selling a technical capability (testing, data, automation) to a non-technical buyer, invest disproportionately in named case studies - they do the technical vouching a sales rep can't. If your category has an entrenched, broad-suite incumbent, don't compete on breadth; compete on speed-to-value and pricing transparency for the segment the incumbent's suite pricing prices out.
gc
Market Context
| MARKET INTELLIGENCE
THE STANDARD: In a fragmented category you rarely need to beat the largest incumbent on features — you need to be the obvious faster-to-buy option for the segment they under-serve.
RULE 1 — DEPLOYMENT SPEED IS THE MID-MARKET'S BUYING CRITERION. Enterprise experience suites assume a dedicated optimisation team; a mid-market marketing lead has none.
RULE 2 — BEING PART OF A LARGER SUITE IS THE INCUMBENT'S STRENGTH AND ITS DRAG. Buyers who don't want the whole platform are structurally unreachable for them.
RULE 3 — EXPERIMENTATION VOLUME IS THE RETENTION METRIC. Tools that don't produce a steady cadence of tests are cancelled at the first budget review.
RULE 4 — THE CATEGORY IS RE-FORMING AROUND AI PERSONALISATION, WHICH FAVOURS DATA SCALE. Independent optimisation tools must attach to a data asset or become a feature.
MARKET TYPE: Fragmented Market (conversion optimisation and experimentation).
| MARKET ENTRY PLAYBOOK
THE STANDARD: A CONSULTING CLIENT BASE IS A LEGITIMATE FIRST MARKET FOR A PRODUCT PIVOT — and self-financing that pivot proves demand before any investor sees it.
RULE 1 — CONVERT SERVICE CLIENTS INTO SOFTWARE CUSTOMERS BEFORE RAISING.
Existing relationships de-risk the pivot and produce reference customers the first sales hire will need.
RULE 2 — RAISE ONLY ONCE THE DIRECT-SOLD BASE IS PROVEN.
Capital arriving after product-market fit funds scale rather than discovery, and the terms reflect it.
RULE 3 — EXPERIMENTATION TOOLS MUST PROVE UPLIFT IN THE CUSTOMER'S OWN REVENUE REPORTING.
Tested outcomes measured anywhere else will not survive a budget review.
How to enter
| FOOTHOLD STRATEGY PLAYBOOK
THE STANDARD: Quantify in the buyer's own financial units — iterations per budget, days off the decision. Specificity converts a deal-driven, sceptical buyer.
SEQUENCE:
1. Attack the analysis bottleneck that gates a financial decision.
2. Express ROI in deal economics, not software language.
3. Demo live, because real-time generation is the differentiator and can't be conveyed in a deck.
WORKED: Real-time capability incumbents can't match, framed in the language the buyer already uses.
CAUTION:
1. IF THE LIVE DEMO IS THE PITCH, IT MUST NEVER FAIL. A crash in front of a client undercuts a speed claim far more than a slow tool's bug ever would.
2. DEAL-DEPENDENT DEMAND COLLAPSES WHEN THE UNDERLYING TRANSACTIONS STOP.
How to get the first strong position
MARKET PATTERNS & PLAYBOOK
3
MONEY
money rev pri
REVENUE MODEL
Subscription
PRICING MODEL
Tiered Pricing, Subscription Discount Pricing, Trial Pricing
WHY THEY WON
Multi-year enterprise contracts, custom-quoted by sales rather than published, scaling with which modules a client turns on (core testing vs. the full suite including feature management and AI personalization) and site traffic volume. Estimated revenue in the $25M-$50M range on roughly 300 employees implies an average contract size in the tens of thousands of dollars - consistent with a mid-market/enterprise, not self-serve, deal profile.
Tiers scale by module (testing-only vs. full suite) and by traffic/seat count; the top tier is custom-quoted rather than listed on a public rate card, keeping enterprise pricing anchored to deal size - the SMB/Shopify tier, by contrast, is closer to a fixed published rate.
TARGET AUDIENCE
CUSTOMER BUYING BEHAVIOUR
tg cb
Enterprise marketing and product teams at consumer brands (retail, luxury, e-commerce) buying measurable conversion lift; digital/UX leads at mid-market e-commerce companies buying self-serve testing through the Shopify app.
Enterprise logos: committee-led, multi-stakeholder (marketing, engineering, legal) sales cycles measured in months, triggered by a specific, quantifiable conversion problem. SMB/Shopify tier: self-serve, trial-first, credit-card checkout with no sales conversation.
| PRICING INTELLIGENCE
What makes this model effective & make customers pay
Experimentation software is priced on traffic tested, and justified by conversion lift the customer measures themselves.
RULE 1 — PRICE ON TESTED TRAFFIC OR SESSIONS, WHICH TRACKS BOTH YOUR COST AND THEIR SCALE.
Larger sites derive proportionally more from the same tests.
RULE 2 — CONVERSION LIFT IS SELF-EVIDENT ROI AND ENDS THE PRICE CONVERSATION.
A small percentage improvement on meaningful e-commerce revenue exceeds any subscription. The customer computes it in their own analytics.
RULE 3 — THE FREE OPEN-SOURCE AND PLATFORM-NATIVE ALTERNATIVES SET YOUR FLOOR.
When adequate testing is bundled into analytics suites, differentiation must be personalisation depth and governance, not the test itself.
RULE 4 — TRIALS ARE ESSENTIAL BECAUSE VALUE CANNOT BE DEMONSTRATED WITHOUT THE CUSTOMER'S OWN TRAFFIC.
No demo proves a lift. The proof requires their data.
A digital team is buying the ability to stop arguing about design opinions and settle it with evidence. Removing internal debate is worth more than the lift itself in large organisations, because it converts a political problem into a measurement one.
PRICE & REVENUE
| Revenue Risk - The biggest threat to revenue stability
Custom-quoted multi-year contracts in the tens of thousands mean a mid-market sales motion where every loss is material and every renewal is a full negotiation.
Pricing on traffic volume ties revenue to customer web activity at the point AI search is reducing it.
Experimentation budgets are discretionary, require in-house expertise to use, and are cut when marketing teams shrink.
The category consolidated sharply after a major incumbent exited; that removes a competitor and signals the category's standalone economics.
Estimated $25-50M revenue on ~300 employees implies contract values in the tens of thousands — a mid-market, not self-serve, profile.
Where the model can break
4
MOTION
Instagram: http://www.instagram.com/abtasty, Facebook: http://www.facebook.com/abtasty, Twitter: http://www.twitter.com/abtasty, LinkedIn: http://www.linkedin.com/company/abtasty
GROWTH EXPANSION MODEL
COMPETITIVE STRATEGY
motion ge cs
Geographic Expansion, Product Line Expansion
HOW THEY EXPAND
The sequence tracks funding: UK, Germany, and Spain offices opened after the 2016 Series A; US and Singapore offices followed the 2018 Series B; the 2020 Series C ($40M) then funded product-line expansion into feature management (Flagship) and AI-driven personalization, moving the company from a single-purpose testing tool toward a broader experience-optimization platform.
Differentiation, Focus Strategy
HOW THEY COMPETE
Rather than attacking Optimizely or Adobe Target head-on for the largest global enterprises, AB Tasty focused on mid-to-large brands wanting an easier-to-deploy alternative, differentiating on implementation speed and support responsiveness rather than trying to out-enterprise vendors already entrenched at the largest accounts.
GROWTH ENGINE
GTM
ge n gtm
Content Flywheel, Partnership Growth
SEO-driven educational content (blogs, guides on CRO) generates inbound leads that direct sales then converts using published, name-brand case studies as proof; a newer Shopify App Store listing adds a secondary self-serve acquisition channel. Where it breaks: the content engine only works as long as AB Tasty keeps outproducing better-funded suite competitors like Adobe on educational content quality and search visibility.
- Content strategy built around educational material on A/B testing, UX optimization, and digital marketing best practice, aimed at generating organic traffic.
- Email nurture campaigns converting leads with personalized recommendations and product updates.
- Active engagement with the digital-marketing community on LinkedIn and Twitter, sharing insights and product news.
- SEO and paid search investment to drive targeted traffic from marketers actively searching for CRO solutions.
SUSTAINING MOATS
Switching Costs, High Customer Lock-In, Brand Power, Technology Advantage (complex enterprise scenarios)
moat
Once a marketing team's test logic, personalization rules, and years of experiment history live inside AB Tasty, ripping it out means rebuilding testing infrastructure from scratch and losing that historical record - that operational entanglement, not brand strength, is what keeps enterprise renewal rates high, and it gets stronger every year more experiments accumulate inside the platform.
| MOAT INTELLIGENCE
THE STANDARD: Experimentation platforms are defended by the accumulated test archive, because the record of what did not work is the part nobody can regenerate.
RULE 1 — THE HISTORY OF FAILED TESTS IS THE INSTITUTIONAL ASSET. Winning variants get shipped and forgotten; knowing which hypotheses have already been disproved is what stops a team repeating three years of work.
RULE 2 — TAGS EMBEDDED IN PRODUCTION CODE MAKE REMOVAL AN ENGINEERING PROJECT. Testing infrastructure deployed across a live site is not a subscription decision — it requires a release, testing and sign-off.
RULE 3 — STATISTICAL CREDIBILITY IS THE PURCHASE CRITERION AT ENTERPRISE SCALE. Once results inform roadmap and budget decisions, methodology defensibility matters more than interface quality.
THE SIGNAL: the category faces a squeeze from feature-flagging tools built for engineers and from analytics suites bundling experimentation free. The defensible position is the organisation that must govern who is allowed to run a test — governance outlives tooling.
Why this company remains defensible
ARR & TAKEAWAY
ARR Journey - what to do at each stage
PRE-$1M ARR — SELL EXPERIMENTATION TO MARKETERS, NOT ENGINEERS
Developer-first testing tools require engineering time marketers do not control. A visual editor that lets a marketing team run tests alone is the wedge.
Land in European e-commerce, where conversion rate is the board metric.
$1–5M ARR — PROVE UPLIFT IN THE CUSTOMER'S OWN REVENUE REPORTING
Experimentation is bought on demonstrated incremental revenue and cancelled without it.
WATCH: tests run per customer per month — the leading churn indicator.
$5–10M ARR — SERVICES AND STRATEGY ARE THE ADOPTION UNLOCK
Most customers lack an experimentation practice. Provide it as a packaged programme, not bespoke consulting.
$10–50M ARR — EXPAND FROM TESTING TO PERSONALISATION AND FEATURE MANAGEMENT
Adjacent products on the same infrastructure raise ACV with the same buyer; acquiring feature-flagging capability was the natural extension.
$50–100M ARR — CONSOLIDATION AND FREE ALTERNATIVES SQUEEZE THE MIDDLE
The category shifted sharply when the largest testing vendor exited free tooling and analytics platforms bundled experimentation. European data-residency and support are the durable differentiators.
NOTE: AB Tasty does not disclose ARR; reported funding varies by source.
$100M+ ARR — NOT CONFIRMED
Rule: experimentation platforms are bought during growth and cut during austerity. Attach to revenue reporting, not to the marketing budget, or you will be first on the list.
COPY PLAYBOOK : What Worked → What Failed → What to Replicate → What to Avoid
THE STANDARD: Convert a services practice into a product only after the workflow has proven itself across real client engagements. Then fund one new geography per round rather than several at once.
SEQUENCE:
1. Run the service until the repeatable workflow is obvious, then productise that.
2. Open one market per funding round so the balance sheet is never stretched across simultaneous cold starts.
3. Build the direct sales team in step with the logos you intend to close.
WORKED: Services-validated product-market fit plus disciplined, sequential geographic expansion.
CAUTION:
1. LOGO COUNT IS NOT PRODUCT-LED GROWTH. A sales-cycle-dependent engine is cash- and headcount-intensive — acquiring big logos without building the team to close and service them stalls the company.
2. SERVICES-ORIGIN COMPANIES CARRY MARGIN AND CULTURE DEBT that takes years to convert into software economics.
bottom of page