Competitor Research: Build Evidence Before You Build a Strategy

Begin with the customer job. Direct competitors solve a similar problem for a similar buyer with a comparable product or business model. Indirect competitors solve the job differently. Substitutes include manual work, internal tools, agencies, spreadsheets, entertainment, or doing nothing. Emerging competitors may be small today but show relevant product, distribution, or advertising movement.

Use a portfolio rather than a flat list. Include a few direct rivals, one or two adjacent alternatives, an emerging player, and a benchmark only when each has a role in the decision. A famous category leader can distort the study if its scale, market, price, and objective differ from yours.

Write inclusion and exclusion rules. For mobile apps, control platform, country, category, business model, audience, maturity, and store presence. For advertising research, add channel, format, objective, destination, and period. Keep these differences visible rather than averaging them away.

Revisit the set when evidence changes. A company can move from adjacent to direct after a launch, enter a market, change pricing, or target a new user. Keep a short reason for every addition and removal so the watchlist does not become a popularity contest.

Turn each question into fields before collection. For positioning, capture source wording, page location, date, intended audience, promise, proof, CTA, and analyst note. For ads, capture advertiser, creative, format, market, channel, observed dates, message structure, offer, and destination. Consistent fields make comparison possible.

Use first-party competitor sources for what a company says and ships: websites, store listings, documentation, release notes, pricing pages, public filings, job posts, and official channels. Use customer sources for experience and language: app-store reviews, support communities, forums, interviews, and your own win-loss research. Use platform libraries and intelligence products for advertising observations.

Every source has bias. Marketing pages present the intended story. Reviews overrepresent people motivated to post. Job listings signal investment but not a shipped capability. Public ad libraries show available ads but may not expose complete delivery. Third-party estimates depend on coverage and method. Preserve the source type next to the claim.

Create an evidence ladder: direct current observation, official historical statement, credible third-party measurement, repeated customer evidence, analyst inference, and unknown. Do not collapse the levels into one confidence score without notes. Date volatile fields such as price, creative, ranking, and feature availability.

Capture short source excerpts only when needed for traceability. Store the URL, title, access date, market, screenshot or record ID, and researcher. A spreadsheet cell saying "strong onboarding" cannot be audited. A dated note naming the exact steps and source can.

Use three connected records. The competitor profile stores stable identity and business-model context. The event log records dated changes such as launches, pricing, releases, campaigns, store events, and positioning shifts. The evidence table stores the source-level observation and confidence.

Normalize vocabulary. Define what counts as a launch, feature, ad variant, offer, product demo, testimonial, new market, or meaningful rank movement. Keep examples and exclusions in a short data dictionary. Without it, two researchers can produce incompatible categories that look precise in a dashboard.

Separate observation and interpretation. "The pricing page added an annual plan" is observable. "The company is prioritizing retention" is an interpretation. "Annual plans improved retention" requires internal or published evidence. This simple distinction prevents many competitor decks from overstating certainty.

Store negative evidence and gaps. Record that a price was unavailable, a supposed feature could not be verified, an ad disappeared, or a market lacked a store listing. Absence may not prove nonexistence, but it changes confidence and tells the next researcher what to check.

Advertising research is one branch of competitor research. Define a market and period, find the relevant advertisers, then compare visible messages and executions. Tag audience cue, problem, promise, mechanism, proof, offer, CTA, format, sequence, language, country, channel, destination, and observed dates.

SocialPeta's verified Advertiser Search supports filters across app type, keyword, new advertisers, channel, countries, OS, developer region, and additional game business fields. Find Competitors can use a keyword, domain, or selected advertiser with industry, channel, user behavior, download and revenue, and sorting filters. These capabilities help construct the comparison set.

Advertiser Detail organizes store and version context with advertising strategy, channels, countries, media, materials, copy, audiences, and other modules where available. Competitor Comparison supports single- and multi-advertiser panels and obtains advertiser information, top creatives, and ad-type distribution. Creative Analysis Reports are a separate route for creating and reviewing competitor creative reports.

Use these observations to identify repeated mechanisms and changes. Do not label visible creative a winner without first-party results. The ad research workflow provides the detailed collection method, while the competitor analysis framework explains how to turn the evidence into a recommendation.

Use a light weekly watchlist for meaningful changes, a monthly synthesis for marketing and product teams, and a deeper quarterly review for strategic decisions. Add event-driven work before market entry, launches, pricing changes, major creative refreshes, or executive planning.

Set change thresholds. Notify the team when a competitor enters a priority market, changes a core offer, launches a relevant workflow, sustains new creative activity, or crosses a defined rank movement. Do not alert on every social post or minor copy edit. Noise teaches people to ignore the system.

Assign owners and expiration dates. Product marketing may own profiles and positioning, growth may own ads and stores, product may own capability verification, sales may own battlecard feedback, and an analyst may own methodology. Every claim should have a review date based on volatility.

Measure use, not database size. Track whether research changed a brief, corrected an assumption, improved a sales response, opened an experiment, or prevented an unsupported decision. Archive fields that nobody uses. A smaller maintained system is more credible than an impressive stale one.

Open with the decision and scope. Name the customer, market, period, competitor set, unit of analysis, and exclusions. A reader should know immediately whether the material can answer their question. Include the collection date because prices, pages, ads, and ranks change.

Provide a compact evidence table before commentary. Each row should contain the observation, entity, source type, URL or record, date, geography, confidence, and limitation. Put interpretations in a separate column. Link the raw record when access and rights allow it.

List conflicts and missing evidence. If a pricing page and sales document disagree, preserve both. If a feature is announced but cannot be verified in the product, label its availability unknown. If an ad was observed but its destination was inaccessible, narrow the claim.

Close with analysis questions, not conclusions the research owner was not asked to make. Examples include whether a new market position is sustained, whether several creative variants share one proof mechanism, or whether a product release addresses a repeated review complaint. State what additional evidence would answer each question.

Include maintenance instructions: volatile fields, owners, alert thresholds, expiration dates, and the next scheduled refresh. This prevents a one-time project from becoming an undated source quoted months later as current truth.

Prepare different views from the same evidence base. Product teams may need workflow and release detail. Growth needs channels, stores, ads, and destinations. Sales needs objection evidence. Executives need the decision, uncertainty, and consequence.

Protect sensitive and licensed material. Define retention and access for downloaded ads, interviews, sales notes, screenshots, and purchased datasets. Public availability does not remove privacy, trademark, copyright, or contractual duties.

Ask a second researcher to reproduce a small sample. If they cannot find the source, apply the taxonomy, or reach the same observation, repair the method before expanding the dataset.