How to do a competitive analysis for a startup idea.

Startup competitive analysis are useful only when they survive contact with the evidence. This guide builds the framework that does.

The most common competitive analysis mistake founders make is treating it as a features exercise. They list their top five competitors, build a matrix of features, mark the cells where they win, and conclude that their product is better. The exercise misses what actually drives outcomes: structural moats, not feature parity. The matrix tells them almost nothing useful about whether they can actually win in the market.

A competitive analysis that informs decisions answers different questions. Who is the current solution for the buyer you are targeting, and why did they choose it? What do they complain about, and how often do those complaints surface? What would it take for them to switch? These questions are about buyer behavior and switching economics, not about feature counts.

Every buyer has a current solution. Before you research any named competitor, research what your target buyer is doing today to solve the problem your product addresses.

For many B2B markets, the current solution is not a software product. According to Gartner's 2023 Software Adoption Survey, 44 percent of business workflows at mid-market companies (100 to 1,000 employees) are still primarily managed through spreadsheets and manual processes, despite the availability of purpose-built software. The incumbent in that market is not a software company. It is Microsoft Excel and the two employees who maintain the spreadsheet.

This matters because the switching cost structure is completely different. Switching from a spreadsheet to your product is a decision about time, training, and perceived risk. Switching from an incumbent software product to yours involves data migration, integration re-work, contract termination costs, and the political capital required to justify replacing a tool someone else championed.

Map the current solution landscape before you map the named competitors. For each current solution type (spreadsheet, consultant, legacy software, adjacent product), document: who in the organization controls the current solution, how much it costs (license plus labor), how painful the known failure modes are, and how often those failure modes occur.

A switching cost model estimates the all-in cost for a buyer to move from their current solution to your product. It should include:

Direct costs: data migration time and cost, integration engineering work, and contract termination penalties if they are leaving a software product.

Indirect costs: the time required to train staff on the new tool, the productivity loss during the transition period, and the political cost of getting internal approval for the switch.

Opportunity costs: the decisions that get delayed while the team is focused on the migration.

According to a 2022 study published in the Journal of Marketing Research analyzing 412 enterprise software switching decisions, the median switching cost for a mid-market SaaS tool with an annual contract value between $10,000 and $50,000 was $28,400 when all direct, indirect, and opportunity costs were counted. The same study found that buyers underestimate their switching costs by an average of 34 percent when asked to self-report.

Your product needs to deliver enough value over a sufficiently short payback period that the switching cost is worth bearing. If a buyer's all-in switching cost is $28,000 and your product saves them $9,000 per year, the rational payback period is more than three years. Most buyers will not make that switch unless they are in active pain with the current solution.

Product review sites are an underused competitive intelligence source. G2, Capterra, and TrustRadius contain thousands of structured buyer reviews that describe exactly what buyers value, what they complain about, and what caused them to switch.

For each named competitor, read the 50 most recent reviews on the platform with the most volume. Tag each review by theme: positive themes (what buyers value), negative themes (what buyers complain about), and switching themes (what caused them to switch from their previous solution or evaluate alternatives).

After 50 reviews per competitor, patterns emerge. The negative themes that appear in 30 percent or more of reviews represent genuine pain points that the market is not satisfied with. Those are the gaps worth building into. The positive themes that appear across multiple competitors represent table-stakes features that every entrant needs to have. Ignore them as differentiators: they are baseline requirements.

Job postings are real-time primary research on competitor strategy. When a competitor posts for an enterprise sales role, they are moving upmarket. When they post for ten engineering roles focused on a specific feature area, they are about to ship something significant in that area. When they stop posting for customer success roles, they may be reducing investment in retention.

Set up job posting alerts for your top three to five competitors on LinkedIn and Indeed. Review them monthly. Over three to four months, the pattern of hiring reveals more about competitor strategy than any press release or product blog.

Job postings also reveal the tools competitors use internally. A posting that requires experience with Salesforce, Zendesk, and Amplitude tells you about their sales, support, and analytics stack. That information is useful for estimating their infrastructure costs and understanding their operational model.

After mapping current solutions, building a switching cost model, analyzing product reviews, and reading job postings, you should be able to articulate the defensible gap: the combination of buyer segment, problem definition, and price point where no current solution is adequately serving the market.

A defensible gap has three components. First, it is large enough to support a real business: enough buyers, at a high enough price point, with a large enough share of their spend you can capture. Second, it is real and not just underserved: buyers are actively spending money or time on a workaround that your product would replace. Third, it is defensible: there is a reason the gap exists, and that reason makes it hard for the obvious incumbents to close it.

The reason a gap exists might be technical (the problem requires a capability the incumbents have not built), go-to-market (the buyer segment is too small for incumbents to prioritize but large enough for a focused startup), or structural (a regulation or market shift created a new need that the incumbents have not yet addressed).

Competitive analysis ends here: with a specific, named gap, the evidence that it exists, and a clear articulation of why it is defensible. Everything beyond that is execution.

Most competitive analyses focus exclusively on direct competitors: the other startups building roughly the same thing. The map that matters for a serious founder also includes substitutes. A spreadsheet. An in-house build by the customer’s engineering team. An agency or consultant. The do-nothing baseline where the customer accepts the pain because the pain is bearable.

Substitutes typically have lower switching cost than direct competitors and serve a meaningful share of the customer base. Ignoring them is the most common cause of TAM inflation. Steve Blank’s customer development frame is direct: "What are they doing today?" is the first question, and the answer is almost never "nothing."

Each competitor and substitute should be scored on six dimensions. Feature parity: how complete is their feature set against the buyer’s actual workflow? Distribution reach: how many target buyers can they reach today through their existing channels? Capital position: how much runway do they have at current burn? Hiring velocity: are they shipping faster or slower than 90 days ago? Integration surface: how embedded are they in the customer’s adjacent tools? Switching cost: what would the customer have to give up to leave them?

The output is a 6-axis radar chart per competitor, not a feature checklist. The shape of the radar tells you where to attack. A competitor that scores high on feature parity but low on integration surface is vulnerable on workflow lock-in. A competitor with low capital position is vulnerable on time. A competitor with low hiring velocity is vulnerable on roadmap speed. The shape suggests the wedge.

The most diagnostic competitive signal is recent shipping cadence. Pull the changelog, the public roadmap, and engineering blog posts from the last 90 days. Count the meaningful shipments. A competitor shipping 6 to 8 features per 90 days has internal momentum. A competitor shipping 1 to 2 has hit organizational drag, regardless of capital position. The drag is the opening.

Hiring is the leading indicator. A competitor that added three senior engineers and two PMs in the last 90 days will ship faster in the next 90. A competitor that has been hiring sales reps but not engineers is signaling a different bet. a16z’s writing on hiring as a leading indicator of strategic direction is useful here. Track public LinkedIn job posts and team-page updates monthly.

The do-nothing competitor wins more often than founders admit. A 30 percent improvement in workflow speed is a strong story until you compare it to the zero-effort, zero-budget, zero-risk option of changing nothing. The do-nothing baseline should be scored on the same six axes as the named competitors. It typically wins on capital cost and switching cost, and loses on feature parity. The wedge has to overcome that asymmetry to make the sale.

Founders who win against do-nothing typically do one of three things: solve a problem that has gotten meaningfully worse recently, ride a behavioral or platform shift that makes the old solution untenable, or attach to a budget that is already approved. None of those are guaranteed in a generic category; they are specific to the wedge and the moment.

Verdikt’s methodology for competitive maps explicitly tests the do-nothing baseline and scores substitutes alongside direct competitors. The output is the competitive shape over the next 6 months, not a snapshot of the current feature gap.