SpyderBot Documentation
Change your measurement scope without misreading progress
Your products, markets, or customer questions have changed. Updating the measurement is appropriate, but a new result may describe a new scope rather than improved or declining performance.
Your outcome: an updated scope with a clear transition record and an appropriate basis for later comparison.
1. Identify what you are editing
| Object | What to confirm before changing it |
|---|---|
| Context | Which product, audience, market, and measurement scope it represents |
| Prompt Universe selection | Which customer needs you want to investigate or observe |
| Brand Insights Prompt Set | The questions intended for that report |
| Benchmark or observation program | The recurring scope and existing comparison history |
Do not assume that editing one object automatically changes all the others. Review the actual destination scope before starting an analysis or recurring observation.
2. Review what changes the benchmark and what only changes your view
When the business changes, start with the benchmark's proposed scope and the decisions you still need it to support. Review Core and Supplemental membership and the competitor set. Use the product's displayed measurement definitions: a Supplemental investigation question is not automatically part of a Core headline aggregate.
| Proposed change | What to check |
|---|---|
| Select a different existing competitor for display | This is a view choice; keep the selected identity consistent across the series you interpret. It does not itself create a new benchmark version. |
| Filter to one AI system | This narrows the viewed scope without rewriting stored all-system results. Missing system evidence remains missing. |
| Add/remove a benchmark question or change Core/Supplemental membership | Review how the actual membership change affects future measurement and comparison history. |
| Change the benchmark's competitor membership | Review the measurement update separately from selecting an existing comparator in the view. |
| Change observation cadence | Review actual observation dates and usage; do not assume a regular historical series was backfilled. |
| Introduce a materially different product, audience, or market | Decide whether a separate Context is needed. |
Keep questions that still matter even when their results are unfavorable. Retire or replace questions because the business need changed, with a recorded reason. If the benchmark already fits, leave it unchanged.
3. Decide whether the business question remains the same
A materially different product, market, or audience often needs a distinct Context. Select or create the appropriate Context through the available controls rather than silently treating a new product as the old measurement.
For a relevant correction or question change, check the product's actual effect on the selection and future observations. Even a wording change may affect the answer. Do not assume historical results will be regenerated or remain directly comparable.
4. Preserve the transition
Keep the last relevant observation reference, the change date, and a short reason. For example: "Added the new travel range; future observations cover commuting and travel, while earlier observations covered commuting only."
Before recurring execution, confirm that the intended questions and Context are in the program. Use a supported comparable view where available. Otherwise mark the new period as a baseline for the updated scope.
5. Explain the next result correctly
If the updated panel includes different customer needs, report its result as such. Do not remove difficult questions merely to raise the headline result. When the product supports comparison within a stable common scope, state that scope explicitly; do not invent a manual aggregate that appears to be an official metric.
You are done when the current scope matches the business question and anyone reading the results can identify what changed and which comparisons remain valid.