top of page
Won product-research category leadership by refusing to treat user testing as a specialist researcher's job — two software developers, frustrated that their prior startup's 2,000-person waiting list had no way to test a prototype at scale, built a tool explicitly meant to let any product manager, marketer, or designer run rapid tests themselves, betting that democratizing research mattered more than serving trained researchers better.
1
MODEL
BUSINESS MODEL
SaaS
model bm
HOW THEY BUILT IT
- Founded 2018 in Paris by Jonathan Widawski and Thomas Mary, both software developers by trade, who built Maze after experiencing firsthand the difficulty of validating a prototype at scale with their previous startup Pingg (a messaging app for gamers) — despite having a waiting list of 2,000 people willing to test the product, they had no platform to run that testing efficiently.
- Positioned explicitly against the traditional model of user research (lengthy, expensive one-on-one interviews requiring specialized researcher expertise), building Maze specifically to empower entire product teams — designers, product managers, and marketers alike — to run quantifiable tests directly from prototyping tools like Figma, Sketch, and Adobe XD.
- Raised a $15 million Series A (2021, led by Emergence Capital) after growing monthly recurring revenue 600% in the prior 12 months to reach $1.5 million ARR and 40,000 companies using the platform, then a $40 million Series B (2022) bringing total funding to $60 million, backed by strategic investors including Atlassian Ventures and Zoom.
- Evolved from a pure prototype-testing tool into what it now calls a 'continuous product discovery platform,' adding features like Maze Discovery (testing without needing a prototype at all), AI-moderated interviews, and an MCP server for AI agent integration, reflecting a broader repositioning as research needs matured beyond a single point-in-time test.
HOW TO ARCHITECT IT
1. If you personally experienced a specific gap during a prior venture (no scalable way to test a prototype despite having eager users ready to test it), consider building the tool you wished you'd had — this founder-market fit gives you genuine conviction and specificity that outside founders lack.
2. Explicitly design your product for the non-specialist user (product managers, marketers, designers) rather than the trained researcher your competitors serve, since democratizing a previously specialized function can expand your addressable market well beyond dedicated research teams.
3. Recruit strategic investors who are also potential enterprise customers or embedded-workflow partners (Atlassian, Zoom) in addition to financial VCs, since these relationships can accelerate both product integration and enterprise credibility.
DISTRIBUTION MODEL
Self-Serve Website, Content Distribution, Platform Integrations
dm
HOW THEY OPERATIONALIZED
Distributed via self-serve trial sign-up reinforced by deep integrations with popular design tools (Figma, Sketch, Adobe XD, InVision, Marvel) that let users import prototypes directly, plus content marketing (design/UX research thought leadership) and a formal partner program with education platform Reforge.
HOW TO REPLICATE WHAT WORKED
What worked: building specifically for the non-specialist user (product managers, marketers, designers) rather than trained researchers, expanding the addressable market beyond dedicated research teams to entire product organizations. Trap if copied blindly: user research and testing tools have become a genuinely crowded category (UserTesting, UserZoom which itself raised $100 million, Pulse Labs) — a founder entering this space today needs a clear differentiation beyond 'democratizing testing,' since that positioning alone is no longer novel as multiple well-funded competitors have adopted similar messaging.
| PATTERNS OF THIS MODEL
PATTERNS IN DEMOCRATISING A SPECIALIST FUNCTION:
1. BUILD THE TOOL YOU NEEDED IN A PRIOR VENTURE. Specific unmet need produces conviction and detail that outside founders cannot replicate.
2. DESIGN EXPLICITLY FOR THE NON-SPECIALIST RATHER THAN THE TRAINED PRACTITIONER YOUR COMPETITORS SERVE. Democratising a specialist function expands the market well beyond dedicated teams.
3. RECRUIT STRATEGIC INVESTORS WHO ARE ALSO POTENTIAL ENTERPRISE CUSTOMERS OR WORKFLOW PARTNERS, accelerating both integration and credibility.
4. AS THE CATEGORY MATURES, MOVE FROM A POINT-IN-TIME TOOL TO A CONTINUOUS PROCESS. One-off usage produces churn; continuous practice produces renewal.
What companies with this model reveal
| OPPORTUNITY INTELLIGENCE
GOLDMINE 1 — BUILD THE TOOL YOUR PRIOR VENTURE NEEDED.
Standard: the founders had 2,000 people willing to test their previous product and no platform to run it at scale. A specific, personally experienced gap gives conviction and precision a market-sizing exercise cannot.
GOLDMINE 2 — DESIGN FOR THE NON-SPECIALIST, NOT THE RESEARCHER.
Standard: enabling designers, PMs and marketers to run quantifiable tests from Figma, Sketch and Adobe XD expands the market far beyond dedicated research teams — the segment competitors were fighting over.
GOLDMINE 3 — RECRUIT STRATEGIC INVESTORS WHO ARE ALSO WORKFLOW PARTNERS.
Standard: Atlassian Ventures and Zoom alongside Emergence accelerate integration and enterprise credibility together.
THE PIT — DEMOCRATISING RESEARCH MEANS SELLING TO PEOPLE WITH NO RESEARCH BUDGET.
$60M raised across 2021 and 2022 priced a design-tooling peak, and the buyer — product and design teams — is exactly where headcount contracted hardest from 2023.
THE SECOND PIT — PROTOTYPE TESTING DEPENDS ON DESIGN TOOLS THAT CAN SHIP TESTING NATIVELY.
Figma is both the distribution channel and the roadmap risk.
MOVE WITH CAUTION — AI-MODERATED INTERVIEWS AND SYNTHETIC USERS ATTACK THE PANEL ECONOMICS UNDERNEATH THE PRODUCT.
Untapped Business Model / Gaps / Goldmines / Pits
Patterns & Insights
2
MARKET
mkt mt es
MARKET TYPE
Fragmented Market
WHY THEY WON
User research and testing software is fragmented among many competitors (UserTesting, UserZoom, Pulse Labs) serving different segments of the research value chain, with over 70% of companies (per Design Tools Survey data cited by Maze) still not using any dedicated testing tool at all, relying instead on lengthy one-on-one interviews. Maze won share specifically by targeting that large underserved segment of non-adopters. Transferable principle: even in a category with several established, well-funded competitors, a large pool of non-adopters relying on manual, unscalable processes can represent a bigger opportunity than competing directly for existing tool-users.
ENTRY STRATEGY
Greenfield Entry
EXECUTION
Maze entered directly via self-serve trial sign-up targeting product design teams, the standard entry mode for a founder-led product-led-growth SaaS startup building from the founders' own direct experience with the underlying problem at their prior startup.
FOOTHOLD STRATEGY
fs
Beachhead Strategy
The beachhead was individual product designers needing quick, quantifiable usability feedback on prototypes before committing engineering resources — a reachable segment given deep integrations with the design tools (Figma, Sketch) designers already used daily. From there, Maze expanded to include product managers, marketers, and eventually entire cross-functional product teams as 'Maze Discovery' broadened the use case beyond pure prototype testing.
GROWTH CAMPAIGN
CAMPAIGNS THAT WORKED
Deep integrations with major design tools (Figma, Sketch, Adobe XD, InVision, Marvel) reducing adoption friction for designers already using those platforms daily; the Series A funding (2021) following 600% MRR growth in the prior 12 months; launch of Maze Discovery, extending testing capability to before a prototype even exists; the Series B (2022), backed by strategic investors Atlassian and Zoom, funding new use-case expansion and workflow integrations; continuous AI feature investment (AI study builder, AI-moderated interviews, AI-powered theme analysis) keeping pace with the broader shift toward AI-assisted research.
KEY LEARNING
If you're evaluating a category with several established, well-funded competitors, look for a large pool of potential customers still relying on manual, unscalable processes (in Maze's case, the 70%+ of companies not using any dedicated testing tool) rather than assuming you must compete directly for existing tool-adopters — that underserved majority can represent a larger opportunity than incremental share-taking from competitors.
gc
Market Context
| MARKET INTELLIGENCE
THE STANDARD: A large pool of non-adopters relying on manual processes can be a bigger opportunity than competing for existing tool users.
RULE 1 — COUNT THE NON-CONSUMERS BEFORE COUNTING MARKET SHARE. Most companies run no dedicated testing at all, which is a larger market than the one competitors fight over.
RULE 2 — SPEED IS WHAT CONVERTS NON-ADOPTION. Teams skip research because it takes weeks; results in hours changes the decision, not the tool preference.
RULE 3 — SELF-SERVE IS REQUIRED TO REACH NON-ADOPTERS. Anyone requiring a sales conversation stays in the non-adopting majority.
RULE 4 — FAST TESTING RISKS SHALLOW RESEARCH, WHICH IS THE CATEGORY'S CREDIBILITY PROBLEM. Depth and rigour must follow speed or the discipline dismisses the tool.
MARKET TYPE: Fragmented Market (user research and testing).
| MARKET ENTRY PLAYBOOK
THE STANDARD: FOUNDERS SOLVING A PROBLEM FROM THEIR PREVIOUS STARTUP ENTER WITH A VALIDATED HYPOTHESIS AND A KNOWN AUDIENCE.
RULE 1 — MAKE THE EXPENSIVE RESEARCH STEP FAST AND CHEAP.
Usability testing was slow and specialist; automating it lets product teams test continuously rather than occasionally.
RULE 2 — INTEGRATE WITH THE DESIGN TOOL WHERE THE WORK ALREADY LIVES.
Testing a prototype where it was created removes the export step that kills adoption.
RULE 3 — RESEARCH BUDGETS ARE DISCRETIONARY AND CUT IN DOWNTURNS.
Tie the product to shipping decisions, not to research maturity.
How to enter
| FOOTHOLD STRATEGY PLAYBOOK
THE STANDARD: Attach to the tools your users already live in and become part of an existing habit.
RULE 1 — TARGET THE DECISION MADE BEFORE ENGINEERING SPEND IS COMMITTED. Validating a prototype prevents wasted development, which is a cost the product team already fears.
RULE 2 — DEEP INTEGRATION WITH THE DESIGN TOOL IS THE DISTRIBUTION CHANNEL. Being present where the work already happens removes the adoption decision.
RULE 3 — QUANTIFYING QUALITATIVE WORK IS WHAT UNLOCKS THE BUDGET. Designers need numbers to defend decisions to stakeholders.
RULE 4 — EXPANDING FROM DESIGNERS TO THE WHOLE PRODUCT TEAM BROADENS THE MARKET AND WEAKENS THE WEDGE. The specific claim must survive the broader positioning.
How to get the first strong position
MARKET PATTERNS & PLAYBOOK
3
MONEY
money rev pri
REVENUE MODEL
Subscription
PRICING MODEL
Freemium, Tiered Pricing
WHY THEY WON
Tiered SaaS subscription priced by number of monthly tests/studies and team size, scaling from individual/small-team plans to enterprise plans with advanced participant recruitment, AI moderation, and workflow integration features.
A free or limited-usage entry tier lets individual designers test the product's value before committing, with paid tiers scaling by testing volume and feature depth (AI moderation, participant recruitment, advanced reporting), targeting product teams who evaluate cost against reduced risk of building the wrong feature.
TARGET AUDIENCE
CUSTOMER BUYING BEHAVIOUR
tg cb
Individual product designers (buying quick, quantifiable prototype testing integrated with Figma/Sketch); product managers and marketers (buying broader concept and messaging validation beyond pure UI testing); enterprise product organizations (buying continuous, cross-functional research infrastructure via Maze Discovery and AI-powered features).
Self-serve trial-first for individual designers and small teams, increasingly sales-assisted for larger enterprise deployments needing broader organizational rollout and participant recruitment infrastructure, a purchase decision typically triggered by a specific product decision requiring validation.
| PRICING INTELLIGENCE
What makes this model effective & make customers pay
Research tooling is priced per study and per participant, because the buyer's constraint is research they cannot otherwise run.
RULE 1 — ANCHOR TO THE COST OF A TRADITIONAL RESEARCH STUDY, INCLUDING RECRUITMENT AND MODERATION.
Agency and moderated research costs are large and slow. Unmoderated testing at a fraction of it reframes the purchase.
RULE 2 — PARTICIPANT RECRUITMENT IS THE MARGIN AND THE OPERATIONAL BURDEN.
Sourcing qualified respondents is where cost and differentiation both sit.
RULE 3 — SPEED CHANGES BEHAVIOUR: FAST RESEARCH MEANS MORE RESEARCH.
When a study takes hours rather than weeks, teams test decisions they previously guessed at. Volume, not cost per study, is the value.
RULE 4 — RESEARCH BUDGETS ARE DISCRETIONARY AND CUT EARLY.
Positioning as risk reduction on expensive builds is more durable than positioning as insight.
A product team is buying evidence before committing engineering time. Where the alternative is building the wrong thing, research prices against development cost — a far larger number than any research budget.
PRICE & REVENUE
| Revenue Risk - The biggest threat to revenue stability
Pricing by monthly tests and team size ties revenue to research activity, which is discretionary and among the first functions cut.
User research teams have been disproportionately affected by tech-sector restructuring — the buyer population itself shrank.
AI-moderated research and synthetic users attack the core value proposition from both directions: cheaper research and questionable validity.
Participant recruitment is a marketplace cost inside a software business.
No current ARR published; last raised $40M Series B (2021).
Where the model can break
4
MOTION
GROWTH EXPANSION MODEL
COMPETITIVE STRATEGY
motion ge cs
Product Line Expansion
HOW THEY EXPAND
Maze expanded from core prototype usability testing into Maze Discovery (testing without a prototype), Reach (participant recruitment and email campaigns), Clips (video and screen recording capture), AI-moderated interviews, and an MCP server for AI agent integration, sequenced to progressively cover the entire product discovery lifecycle rather than a single point-in-time usability test.
Differentiation
HOW THEY COMPETE
Maze differentiated against research-specialist-focused competitors by explicitly designing for the broader product team (not just trained researchers), a sequencing that expanded its addressable market and let it compete on accessibility and speed rather than research methodology depth alone.
GROWTH ENGINE
GTM
ge n gtm
Product-Led Growth, Platform Integrations
Growth compounds through deep integrations with popular design tools that make Maze the natural next step for designers already working in Figma or Sketch, combined with strong word-of-mouth within the product design community. It would break down if design tools themselves (Figma, in particular) built sufficiently capable native user-testing features, reducing the need for a separate best-of-breed testing platform.
Self-serve trial funnel reinforced by deep design-tool integrations and content marketing targeting product teams researching how to validate ideas faster, extended through strategic partnerships (Atlassian, Zoom) and the Reforge partner program.
SUSTAINING MOATS
Switching Costs, High Customer Lock-In, Brand Power, Technology Advantage (complex enterprise scenarios)
moat
Maze's moat is the switching cost of migrating accumulated research history, participant panels, and integrated design-tool workflows to a competing platform, combined with strong brand recognition within the product design and UX research community built through consistent product-led growth and design-tool ecosystem integration.
| MOAT INTELLIGENCE
THE STANDARD: Making research self-serve for product teams expands the market beyond the specialists who previously gatekept it.
RULE 1 — THE BUYER IS THE PRODUCT MANAGER, NOT THE RESEARCHER. Removing the dependency on a scarce specialist is the entire value proposition, and it is why the category grew from a function into a workflow.
RULE 2 — THE PARTICIPANT PANEL IS THE OPERATIONAL ASSET. Recruiting the right testers quickly is a supply-side capability, not a software feature, and it is what makes the product usable on a deadline.
RULE 3 — ACCUMULATED STUDY HISTORY IS INSTITUTIONAL KNOWLEDGE, because knowing what has already been tested prevents teams repeating research nobody remembers commissioning.
THE SIGNAL: models can now synthesise plausible user feedback, which threatens the volume end of research and not the credibility end. The defensible position is verified human participants — the thing generation cannot supply.
Why this company remains defensible
ARR & TAKEAWAY
ARR Journey - what to do at each stage
PRE-$1M ARR — MAKE USER TESTING FAST ENOUGH FOR A DESIGN SPRINT
Traditional research takes weeks. Testing a prototype with real users in hours changes when research happens rather than how.
Sell to product designers and researchers at software companies, who feel the delay directly.
$1–5M ARR — INTEGRATE WITH THE DESIGN TOOL, NOT WITH THE RESEARCH PROCESS
Launching a test directly from the prototype is the mechanic that removes friction entirely.
WATCH: tests run per account per month.
$5–10M ARR — PARTICIPANT RECRUITMENT IS THE BOTTLENECK
Access to relevant testers is what stops teams from researching. Solving supply is more valuable than the analysis.
$10–50M ARR — DEMOCRATISE RESEARCH BEYOND RESEARCHERS
Researchers are few; product managers and designers are many. Expanding the user base is the growth path.
Reached a reported $40M Series B in 2021.
$50–100M ARR — DESIGN TOOL DEPENDENCY AND AI BOTH COMPRESS THE CATEGORY
Design platforms add testing natively while AI simulates user feedback. Real participant data and research rigour are the defensible remainder.
NOTE: current ARR is not disclosed.
$100M+ ARR — NOT IN EVIDENCE
Rule: making a slow process fast changes when it happens, which expands the market. Then the constraint moves to supply — solve that before someone else does.
COPY PLAYBOOK : What Worked → What Failed → What to Replicate → What to Avoid
THE STANDARD: Building for the non-specialist rather than the trained practitioner expands the addressable market from a dedicated team to an entire organisation.
SEQUENCE:
1. Identify the specialist function whose work others would do if it were accessible.
2. Build for the non-specialist without requiring the specialist's training.
3. Expand from the research team to the whole product organisation.
WORKED: Building for product managers, marketers and designers rather than trained researchers, expanding the buyer from one team to the organisation.
CAUTION:
1. "DEMOCRATISING RESEARCH" IS NO LONGER A NOVEL POSITION — the category is crowded with well-funded competitors using near-identical messaging. Entering today requires differentiation beyond accessibility itself.
2. NON-SPECIALIST USERS PRODUCE LOWER-QUALITY INPUTS, which can undermine the outcome the product promises.
bottom of page