Original research
A taxonomy of competitor website changes, with real examples
A reusable classification system for organizing commercial, product, positioning, proof, conversion, and presentation changes.
A change taxonomy makes a monitoring system easier to tune and a competitive report easier to read. The categories below are designed for real public website evidence; examples should be replaced or expanded with cited observations from the pages being analyzed.
Use categories that map to decisions
Classify a change by its likely decision surface: commercial, positioning, product, proof, conversion, discovery, or presentation. The classification is not a conclusion about intent. It is a way to route the evidence to the people who can evaluate it.
Keep a separate field for confidence and urgency. A product-page claim may be high-confidence evidence of published messaging but low-confidence evidence of product maturity or strategic priority.
| Category | Example evidence | Likely owner |
|---|---|---|
| Commercial | Price, plan, limit | Product and finance |
| Positioning | Audience or category claim | Product marketing |
| Product | Feature or integration page | Product |
| Proof | Customer story or metric | Marketing and sales |
| Conversion | Signup or sales CTA | Growth and sales |
Document examples without overclaiming
For each example, show the page type, the observed text or element, the checked date, and why it belongs in the category. Do not infer a private motive from the label. A “contact sales” CTA is a conversion observation; a claim that the company is abandoning self-serve is a hypothesis.
Use the meaningful-change rubric to rank examples and the facts-versus-interpretation guide to write them safely.
Make the taxonomy evolve
A useful taxonomy is neither universal nor permanent. Add a category when several changes create a recurring decision, merge categories that cannot be distinguished in practice, and keep an “unknown” path for evidence that needs review.
Publish the method and revision date. Readers should be able to understand how an example was classified and how the classification may change as the dataset grows.
Use a taxonomy that supports action
A useful taxonomy describes the changed object and its likely decision context: pricing, packaging, capability, positioning, proof, route to purchase, integration, editorial, legal, or technical availability. Keep 'observed type' separate from 'business implication' so analysts do not encode a conclusion as if it were a fact.
The category should be assigned from visible evidence and stable rules. For example, a new plan table row is packaging evidence; an analyst's claim that it targets enterprise buyers belongs in an interpretation field. This distinction makes summaries safer and supports later review.
| Type | Evidence example | Follow-up |
|---|---|---|
| Commercial | Price, term, or limit changed | Recalculate scenarios |
| Product | Capability or integration added | Check docs and proof |
| Messaging | Audience or outcome reframed | Review positioning |
| Access | CTA or signup path changed | Test buyer journey |
Handle ambiguous and compound edits
One deploy can change a navigation label, a plan limit, and a customer story. Record the atomic observations first, then link them to one scan or release. If a change fits multiple categories, retain the primary category and add a secondary label rather than forcing a false single explanation.
Exclude unstable content such as rotating hero text when repeated rendered samples show that it is not durable. Also distinguish backfilled blog entries from genuinely new entries. The change monitoring guide describes why stabilization and chronology matter.
Version the taxonomy and audit it
Write definitions, examples, exclusions, and confidence rules before labeling a large history. Version the taxonomy when a category changes, and retain the original label so longitudinal reports remain interpretable. Review a sample with more than one analyst and record disagreements.
Measure precision and missed-signal rate against a labeled review set if the taxonomy drives alerts. Treat those figures as a research plan until measured. A taxonomy is successful when it routes evidence to the right decision, not when every edit receives a confident label.
