Attention Score
Relative to its own history
Unavailable
Eligible sources end in different months, so they cannot be combined without mixing dates.
Brand overview
Pick one brand and see how its public attention signals moved together, month by month.
New Balance
Sep 2026 · changes compared with Aug 2026
New Balance scorecard
ExperimentalSentiment Score
Based on news headlines
Update delayed
News collection is older than 3 days or never succeeded.
Oct 3, 2026–Nov 2, 2026
52 usable headlines · 23 publishers · Last update Oct 17, 2026
Tone of collected headlines only. Our feeds don’t capture all news.
Google search change
Search interest only vs last month
-25.3%
Search only—not a combined change
Sep 2026 vs Aug 2026
Online activity and headline tone—not sales or a public opinion survey.
How we calculate this
Attention: each eligible source is ranked against New Balance’s own Apr 2026–Sep 2026 months, with half credit for ties. The score is the equal-weight average of those ranks. At least two sources need six complete, comparable months. Equal weights are a design choice, not a proven measure. Method attention-v1.
Attention Change: percentage change in a fixed-weight activity index—not a change in the bounded Attention Score. Each source is measured against its own reference average. Both months must use the same sources and collection rules. Combined change is unavailable: Eligible sources end in different months, so they cannot be combined without mixing dates.
Sentiment Score (News tone): 100 × (positive headlines + half the neutral headlines) ÷ usable headlines. Needs 20 usable headlines and 3 publishers, plus a fresh collection. Mixed or unscorable items are left out; no social media or surveyed opinions are included.
The scores cover different dates—read them separately. Positive/negative headline labels and rising/cooling activity labels use chosen cut-offs, not statistical significance. More online activity does not mean more approval or purchases. Read the full methodology.
Source breakdown
| Source | Latest month | Value | Rank in own 6 months | Weight | Status |
|---|---|---|---|---|---|
| Google search interest | Sep 2026 | 8.8 | 8 | 0% | Eligible |
| Wikipedia page visits | Oct 2026 | 20,117 | — | 0% | 1 of 6 reference months are missing or incomplete. |
| Review videos about the brand | Oct 2026 | 3 | 50 | 0% | Eligible |
| Videos posted by New Balance | Not used—this is the brand’s own publishing, not public attention. | ||||
News tone sample Oct 3, 2026–Nov 2, 2026: 52 usable headlines from 23 publishers. 19 positive, 29 neutral, 4 negative; 0 left out (mixed or not scorable). Last successful news collection: Oct 17, 2026. Our feeds do not capture all news. Rules tone-v2, relevance-v4.
Google search forecast
ExperimentalA one-month outlook for New Balance’s search interest
Archived forecast. A newer export is needed for the current month.
Oct 2026 forecast
12.0 / 100
Historical-error range
9.0–15.0
Average error in historical tests
1.5 points
Blue: observed historyDashed: forecastShaded: historical-error range
Same as last month · Saved 2026-10-08 · Data through Sep 2026. Range uses the 80th percentile of historical errors; future coverage is unvalidated.
Historical model comparison
| Model | Average absolute error |
|---|---|
| Same as last month (selected) | 1.54 points |
| Average of the last 3 months | 1.77 points |
| Exponential smoothing | 1.66 points |
81 rolling historical tests, Jan 2020–Sep 2026. These tests use a later downloaded history and also select the model, so they can overstate future accuracy. This partial-month forecast was saved after Oct 2026 began. Future actuals await a comparable-scale source.
View Google Trends sourceScores use a separate monthly Google Trends export, so their scale differs from Explore. They measure search interest. How it works
Google search interest
8.8
Interest score (0–100 export scale), monthly average
-25.3% vs previous month
What this shows: How often people searched this brand name on Google in the U.S., relative to the busiest day in its export. Not a search count.
Wikipedia page visits
23,684
Page visits
+22.0% vs previous month
What this shows: Visits to the brand’s English Wikipedia page each month. Reference-page interest, not demand or sales.
Official YouTube uploads
7
Videos
-22.2% vs previous month
What this shows: Videos the brand itself posted on its official YouTube channel. Shows the brand’s own publishing, not consumer demand.
News coverage
Unavailable
Articles
Incomplete month — not fully collected, so no value is shown.
What this shows: Brand-related articles saved from our news feeds. Counts show collected articles, not all news coverage.
How the signals moved together
Each line starts at 100 in May 2026, the first month where every selected signal has a complete, positive value. A value of 120 means 20% higher than that month. This compares movement, not raw volumes, sentiment or sales. This 100 is not the Google Trends 0–100 scale.
Gaps mean a month was missing, incomplete, or collected under a different method; nothing is filled in.
Monthly raw values
| Month | Search interestInterest score (0–100 export scale) | Wikipedia page visitsPage visits | Official uploadsVideos | News coverageArticles |
|---|---|---|---|---|
| Apr 2026 | 25.6 | — | 7 | — |
| May 2026baseline | 19.9 | 24,826 | 5 | — |
| Jun 2026 | 14.9 | 22,114 | 45 | — |
| Jul 2026 | 11.1 | 22,840 | 7 | — |
| Aug 2026 | 11.7 | 19,416 | 9 | — |
| Sep 2026 | 8.8 | 23,684 | 7 | Partial data |
About this data
Google search interest. Google Trends, United States, Web Search, All categories, literal search terms (not Topics), Apr 1–Sep 30, 2026. Each Google Trends comparison export is scaled 0–100 on its own (100 = the busiest day for any term in that export). The monthly score is the average of the daily scores for every day of the month; months with missing days are not shown. Trends values are sampled and rounded, so scores are approximate relative interest, not search counts. Literal brand-name searches can include other meanings (e.g. the animal puma, Ford as a surname) and do not prove purchase intent, sales or sentiment. Imported manually from CSV files downloaded Oct 1, 2026; this source is not refreshed automatically. Months are only compared within one export batch and scale.
Wikipedia page visits. Wikipedia attention = monthly English Wikipedia pageviews by human users (all devices) for each brand's verified article, over the six most recent complete calendar months. It measures reference-page interest only — not overall demand, sales, news coverage or sentiment.
Official YouTube uploads. Owned YouTube uploads = public videos published in each UTC calendar month on one fixed official channel per brand (panel yt-owned-panel-v1), enumerated from the channel's uploads playlist and re-verified with videos.list, deduplicated by video ID. It measures what the brand itself publishes — not all YouTube videos about the brand, not views, not consumer demand. A month is counted only if its scan reached past the month start; otherwise it shows as unavailable, never zero. Deleted or private videos are not counted. Uploads reflect the brand’s own publishing, not consumer demand. Months are only compared under the same channel and panel version.
News coverage. Unique, clearly relevant articles (brand named in the headline; quote pages and uncertain matches left out) collected from GDELT DOC 2.0 and Google News RSS, deduplicated by publisher URL and headline. This is sampled tracked coverage, not all news; feed fetch limits cap each run. Google News dates are publication dates; GDELT dates are first-seen dates. Only fully collected months are charted or compared.
This chart does not combine signals into one score. The separate experimental scorecard above follows its own documented method on the Methodology page. Legacy YouTube search estimates and social data are never shown here. To compare several brands on one signal, use Compare on Explore.