Monitor the smallest set of primary pages that could change a real decision. For most software companies, begin with pricing, a changelog or release feed, product documentation, and the homepage. Add jobs, status, trust, policy, or integration pages only when a named owner can explain what they would do with the change.

What this guide covers

  • the highest-value source types;
  • which decisions each source can inform;
  • how to prioritize sources by directness, volatility, and actionability;
  • two copyable source maps.

Use a decision-led source map

Do not start with an exhaustive crawl of the competitor's domain. Start with the decision and work backward.

DecisionFirst sourceSupporting sourceCommon trap
Compare price/packagepricing or plansterms, help, calculatorrecording only headline price
Track shipped capabilitychangelog/release notesdocumentationtreating every release as strategic
Detect limit/deprecationdocumentationchangelog, statusmissing technical pages not linked in nav
Track positioninghomepage/product pagecustomer storiesinterpreting an A/B test as a pivot
Assess enterprise readinesstrust/security pagepricing, docsassuming a logo proves scope
Watch reliabilitystatus pageincident archivecomparing raw counts across disclosure policies
Investigate investmentjobs pageproduct, partner, docstreating one vacancy as strategy

The EXVIV company directory shows why source coverage differs by company. A company profile is a discovery map; your watchlist should still include only pages tied to a current decision.

Tier 1: direct commercial truth

Pricing and plan pages

Watch these for visible prices, tiers, value metrics, included usage, entitlements, trials, billing terms, and calls to action. GitHub Copilot's plans page is the kind of source that can support current public plan comparisons.

Add the pricing FAQ, calculator, or plan-comparison page when it defines the billable unit more precisely than the marketing page.

Terms and policy pages

Terms can reveal fair-use limits, data restrictions, renewal conditions, regional availability, and policy changes. Monitor only the relevant section or structured field; legal templates produce noisy full-page diffs.

Tier 2: shipped product truth

Changelogs and release notes

Use official dated feeds to track launches, improvements, integrations, and deprecations. Gemini API's changelog is more useful for API lifecycle questions than a homepage capture.

Prioritize entries that change availability, compatibility, packaging, limits, or migration risk. Routine fixes belong in a digest or no alert at all.

Documentation

Documentation often contains the most precise limits and implementation boundaries. Watch:

  • model or feature availability tables;
  • quotas and rate limits;
  • API versions and deprecations;
  • deployment and data-region options;
  • export and ownership terms;
  • integration setup pages.

Documentation trees change frequently. Monitor the exact page or stable section, not the whole documentation site.

Tier 3: market and operating signals

Homepage and product pages

These show the public audience, promise, proof, and CTA. They are strong evidence of a message change and weak evidence of the reason behind it.

Capture the hero, audience labels, proof, and CTA as fields. Recheck a major change across multiple captures or pages before calling it a repositioning.

Status and incident pages

Bolt's status page can provide incident and component records. Save incident start, update, resolution, affected component, and stated impact. Do not reduce the record to “number of incidents.”

Trust and security pages

Use these for certifications, subprocessors, data handling, deployment models, and governance claims. A quote-led page such as Vanta's pricing page can also show which enterprise capabilities are visible while leaving exact commercial terms unknown.

Jobs pages

Track role clusters, locations, seniority, and repeated capability language. One opening is an observation. A durable cluster plus corroborating product or partner changes can support an investment hypothesis.

Integration and partner directories

Monitor a small set of strategically important integrations. A listing proves visibility; detailed setup documentation and reciprocal announcements provide stronger evidence of a working relationship.

Score a source before adding it

Rate each candidate from 1 to 5:

FactorQuestion
DirectnessDoes the company define the fact here?
Decision valueWould a material change trigger a named review?
ComparabilityCan the relevant fields be normalized over time?
StabilityCan region, login, and experiments be controlled?
TimelinessIs the source updated near the underlying change?

Add sources with strong directness, decision value, and comparability. A volatile page can still be useful if normalization is good. A perfectly stable page with no decision value is not.

Source map for a SaaS competitor

Company:
Reason in peer set:
Decision owner:

1. Pricing/plans
   URL:
   Fields: plan, price, period, unit, limits, entitlements, CTA
   Review rule:

2. Changelog
   URL:
   Fields: release, availability, package, deprecation
   Review rule:

3. Documentation
   URL:
   Fields: limits, API, region, migration
   Review rule:

4. Homepage
   URL:
   Fields: audience, promise, proof, CTA
   Review rule:

Optional: status / trust / jobs / integrations
Reason:

Source map for an AI infrastructure vendor

Add these fields to the SaaS map:

  • model and version availability;
  • input, output, cached, and batch rates;
  • context or throughput tiers;
  • regional processing;
  • deprecation and migration dates;
  • service components and incident state;
  • enterprise data controls.

Use the EXVIV market map to bound the category and Evidence Compare to inspect comparable records. The full competitor-monitoring guide explains the baseline and review loop.

Sources and further reading

Method note

Source selection is contextual. A page belongs in the watchlist because it can improve a decision, not merely because it changes. Monitor public pages respectfully and keep private captures and rejected observations outside the public research layer.