Monitoring system
Build market research feeds around signals you can verify
Design market research feeds for continuous monitoring, with clear signal types, review rules, and a log that separates observations from conclusions.
Open the signal feedA market research feed keeps a question open after an initial project ends. Its job is to surface new evidence that may change a category view, product decision, or positioning claim. It is not a complete market model and it should not turn every repeated headline into a trend.
The most reliable setup separates signal types before collection and records observations before interpretation. That makes it possible to trace each conclusion back to the material that caused it.
Split the market into signal feeds
Use only the feeds that can change the decision. The examples show how one onboarding market question can be separated without losing the shared context.
| Signal | What belongs | Example query boundary |
|---|---|---|
| Product change | Launches, packaging, integrations, or workflow changes | B2B onboarding product changes and implementation features |
| Customer problem | Attributed language, support themes, adoption barriers, or buying criteria | Enterprise setup delays, time to value, and onboarding ownership |
| Operating example | A documented process, implementation, or organizational choice | How SaaS teams staff and measure customer onboarding |
| External constraint | Regulation, platform policy, technical standard, or procurement requirement | Security and procurement requirements affecting onboarding tools |
| Category narrative | Definitions and claims used to frame the market | How onboarding platforms describe implementation and adoption |
Record the observation before the conclusion
Write what the source establishes in plain terms: a vendor changed packaging, three named buyers described the same implementation barrier, or a policy now applies to a defined group. Keep the source URL and date beside that observation.
Interpretation comes next. A packaging change may suggest a shift in how a product is sold, but it does not prove a category-wide pricing trend. Several articles may trace back to the same announcement. Separating the observation from the conclusion makes those limits visible.
Use a monitoring log
| Field | What to record |
|---|---|
| Research question | The decision or assumption this feed monitors |
| Observed change | Only what the source directly establishes |
| Signal type | Product change, customer problem, operating example, external constraint, or narrative |
| Source and date | Original URL, publisher, and publication date |
| Confidence | Single source, corroborated, or primary evidence |
| Interpretation | What the observation may change, with the uncertain part qualified |
| Next action | No action, gather more evidence, update a claim, or revisit a decision |
Set the review cadence from the decision
- 01
Weekly for active launches and policy changes
Use a short review when delayed awareness would affect a campaign, product response, or customer conversation.
- 02
Monthly for category and customer-problem monitoring
Group observations before interpretation so a single active news cycle does not dominate the note.
- 03
Quarterly for assumptions that change slowly
Review whether the feed still serves a live decision. Retire it when new items no longer change the market view or when ownership has moved elsewhere.
Read curaft counts as sample descriptions
curaft topic pages calculate source counts, publication dates, and recurring tagged themes from the stories loaded for the current search. The figures describe that result sample. They do not measure market share, customer sentiment, or statistical significance.
Use the snapshot to inspect coverage and identify source concentration. Use original sources, customer evidence, product data, and commercial data for decisions that require a fuller market view. The methodology page documents how the visible feed statistics are calculated.
Retirement rule
Close a feed when its research question no longer maps to an active decision, when another owner has absorbed the monitoring job, or when repeated reviews produce no new evidence worth recording.
Continue the research
Use a live topic for current material or move to the guide that owns the next part of the workflow.