Independent trend intelligence · Data checked 17 November 2026 MethodologyHow we research RSS

Transparency

How we measure a trend

A trend is a direction in attention data, not a prediction. Here is exactly what we calculate, what we do not, and where the error bars are.

Updated 17 November 2026 TrendDocket editorial

Our public trend pages use two independent attention sources: Google Trends search interest and English Wikipedia pageviews. Both measure attention, not sales, revenue, or market share. We treat them as evidence of changing interest, not as a forecast.

Every trend page shows the source, the geography, the time window, and the index definition next to the chart. If a data point is missing, we say so rather than filling the gap with an estimate.

Google Trends: what the 0–100 index means

Google Trends normalises a query's interest to its own peak. A value of 100 marks the highest point in the selected time window; 50 means half that relative interest. The index is not a count of searches, and values from separate queries cannot be compared directly unless they were requested in the same comparison.

We request worldwide, web-search data for the last 12 months, then use complete weeks only. The final partial week is excluded. Our 13-week change compares the latest four complete weeks with the prior 13 complete weeks. Our 12-month change compares the latest four complete weeks with the first four weeks of the window.

These windows answer different questions. The 13-week change is momentum; the 12-month change is direction. A launch spike can make one positive and the other negative. We show both and label them separately.

Wikimedia pageviews: the second signal

We use the Wikimedia Pageviews API for English Wikipedia articles related to a topic. The data is monthly, all-access, user-agents only. We compare a three-month recent window with the prior three months.

Wikipedia attention is not purchase intent. People read an article to understand a topic, check a product category, or follow the news. It is useful as a corroborating signal because it is independent of Google Trends and less sensitive to a single search-results page.

When the two signals diverge, we say so. Wi-Fi 7 is the clearest example in our launch set: search interest fell while reference reading rose. That pattern is consistent with a shift from news curiosity to purchase research.

Status and momentum

<strong>Status</strong> reflects the 12-month direction. Rising means a 12-month change of at least +10%. Declining means −20% or lower. Stable covers the range in between. A Rising signal without usable regional corroboration is labelled Emerging until a second independent source confirms scale.

<strong>Momentum</strong> reflects the latest four complete weeks versus the prior 13 weeks. Accelerating is +5% or better. Cooling is −10% or worse. Steady is in between. This is why a page can say “Rising / Cooling”: the annual direction is up, but the recent quarter is fading.

The <strong>signal score</strong> is a relative ranking, not a probability. It combines the 12-month change (50%), the 13-week change (30%), and the latest four-week index (20%), each min-max normalised across the tracked set and scaled to 0–100. It is useful for ordering a list; it is not a forecast.

What we do not measure

We do not have unit sales, revenue, search volume, market share, or affiliate conversion data in this launch dataset. We do not infer them from search interest. When a page says “attention”, it means attention.

We do not publish confidence intervals for the public signal score. The underlying Google Trends data is a relative index with sampling and normalisation effects, and a single index value is not a precise measurement. Read the direction and the time window, not the decimal place.

Update cadence and corrections

Google Trends and Wikimedia snapshots are refreshed manually for this launch. Pages show the date of the last data check. Product prices are checked separately and are not automatically refreshed.

If a source changes or a calculation is wrong, we correct the page and note the change. The internal audit log records the editor, time, and reason.

References

Sources

  1. Google · Retrieved 2026-11-17
  2. Wikimedia Foundation · Retrieved 2026-11-17

    Monthly pageviews, en.wikipedia, all-access, user agents only.

  3. Google Trends · Retrieved 2026-11-17

    Weekly relative search-interest index, 0–100. Not search volume.

  4. Google Trends · Retrieved 2026-11-17

    Weekly relative search-interest index, 0–100. Not search volume.