Module 6 · Lesson 2/3 · ~10 min read
0/21 lessons read · Your progress is kept in this browser only.
Reading Week-to-Week Trends
When is a weekly SOV change shown, and when is it hidden? Comparability conditions, versions and partial weeks, worked through five representative weeks.
Learning goals
- Lists what a weekly change needs before it's shown: the same question set, competitor set, scope and engine configuration, with all five engines healthy in both weeks.
- Knows that otherwise the change is hidden and the reason is shown.
- Explains why a trend stops being comparable once the competitor or question set changes.
- Doesn't read a partial week as a full one, and tells percentage points apart from percent change.
In this lesson
A single SOV number is a snapshot. The question people actually ask is "what changed since last week?" This lesson covers when that question can be answered honestly.
Weekly SOV and its trend are measured on paid plans; the free scorecard has no SOV (Module 2).
Comparability conditions
Comparing two weeks of SOV only makes sense if both weeks measured the same thing. A week-to-week change is shown only when all of these hold:
- Same question set: the tracked questions and their wording haven't changed.
- Same competitor set: the competitors in the denominator are the same.
- Same scope: same brand, country, language and, where set, the same place, channel or focus.
- Same matching and metric rules: the brand's recognised spellings and the metric definitions are unchanged.
- Same engine configuration.
- All five engines healthy in both weeks: as Module 2 explained, an engine is healthy when at least 95% of its scheduled questions were measured that week.
If any condition fails, the change is hidden and the reason is shown, for example "competitor set changed" or "partial measurement this week".
Versions
Changing a question's wording creates a new question version. Adding or removing a competitor creates a new competitor-set version. Charts mark these change points.
Why? Because the SOV denominator changes. Add a competitor and your share shrinks even if your brand was mentioned exactly as often. That isn't a loss; the universe you're measuring has changed.
Partial weeks
Recall the rule from Module 2:
- 5 of 5 engines healthy: SOV is complete.
- 3 or 4 healthy: SOV is shown as partial, with the engine count (e.g. "partial · 4/5 engines").
- Fewer than 3 healthy: no SOV for that week. On a chart that week is a gap, not a drop to zero.
You can read a partial week on its own, but no change against the previous week is shown. Four engines and five engines don't measure the same thing.
Worked example (representative)
Acme Helpdesk, its competitors and the numbers are fictional.
| Week | Competitor set | Healthy engines | Mentions (Acme / A / B / C) | SOV | Change vs previous week |
|---|---|---|---|---|---|
| 1 | A, B | 5/5 | 6 / 8 / 6 / – | 30% (6 ÷ 20) | First week |
| 2 | A, B | 5/5 | 7 / 8 / 5 / – | 35% (7 ÷ 20) | +5 points |
| 3 | A, B, C (new version) | 5/5 | 7 / 8 / 5 / 9 | 24% (7 ÷ 29) | Hidden: competitor set changed |
| 4 | A, B, C | 4/5 | – | 26% · partial · 4/5 engines | Hidden: partial measurement this week |
| 5 | A, B, C | 5/5 | – | 27% | Hidden: previous week was partial |
How to read it:
- 1 → 2: Nothing changed and both weeks are 5/5, so the change is shown: +5 points. Careful: that doesn't mean "up 5%". Going from 30% to 35% is 5 percentage points; the relative increase is 5 ÷ 30 ≈ 16.7%. When you report it, say which one you mean.
- 2 → 3: Acme Helpdesk is still mentioned 7 times, but adding Competitor C raised the denominator from 20 to 29. SOV appears to "fall" from 35% to 24%. The change is hidden because the two weeks measure different universes.
- 3 → 4: Week 4 is partial. SOV is shown with the engine count, but no change is shown.
- 4 → 5: Week 5 is complete, but the week it would be compared with is partial. The change stays hidden.
Reading rules
- Don't calculate a hidden change yourself. Saying "we lost 11 points" from 35% to 24% treats a competitor-set change as brand performance.
- Don't call one small weekly move a trend. Each measurement is a sample of the answers returned that week. Look across several comparable weeks.
- Keep daily checks separate. On paid plans, daily priority checks on ChatGPT and Google AI Overviews are an early signal. They never feed into weekly SOV and aren't compared with it.
- Change isn't cause. Even when SOV rises between two comparable weeks, that doesn't prove one page change caused it.
Plan your changes
- Batch competitor and question changes instead of spreading them out, so there's a single version switch.
- Note what you changed and when.
- After a change, wait for at least two comparable weeks on the new version before drawing conclusions.
Rules: When SOV is measured and what we don't claim to measure.
Summary
- A weekly change is shown only with the same question set, competitor set, scope, matching and metric rules and engine configuration, with 5/5 engines healthy in both weeks.
- If any condition fails, the change is hidden and the reason is shown.
- Changing the question or competitor set creates a new version. The denominator changes, so it can't be compared with earlier weeks.
- Partial weeks (3–4/5) are shown with the engine count but without a change; below 3 engines there's no SOV, and the chart shows a gap.
- Keep percentage points and relative change apart, don't calculate hidden changes yourself, and don't treat change as cause.