Measure a past release
Get a verdict on something you shipped weeks or months ago, from the history Needlr already keeps.
Something shipped three months ago, and nobody measured it. An impact check can still give you a verdict, as long as Needlr was already counting that KPI before the release.
You can get a verdict on a release from the past: last week, or last spring. Needlr compares the weeks before it with the weeks after, using the history it already keeps.
Think of a height chart
Some families mark their child's height on the kitchen door every month. If you marked it all year, you can see how much they grew last summer.
If you only started marking in September, you can't. The door only shows what was marked.
Needlr's history is that door. It can only look back to the day it started counting your KPI.
How far back you can go
For every KPI it counts, Needlr keeps a daily summary for 13 months. Every account can measure a release up to 395 days back, the full 13 months (plans).
The rule: 4 weeks before
To compare, Needlr needs at least 4 weeks of your KPI from before the release. It can only count a KPI from the day it started counting it. So a KPI you made last week can't measure a release from last month.
Step 3 tells you how much history your KPI has: History from a date, and The earliest release it can measure is a date.
How to measure a past release
- 1On your project page, press + New impact check. Give it a short name and pick Before vs after.
- 2Set the Release date to the day it went live for everyone.
- 3Which needle should it move? Pick a KPI you already have.
- 4How should we test it? Read what step 3 says about the KPI's history. If there's a problem, it says what to do.
- 5Press Start measuring. For a release more than 4 weeks ago, the verdict comes soon after, usually within the hour.
How the comparison works, and what the result means, is in Before vs after.
Old releases use daily summaries
Needlr keeps every single event for 90 days. After that, it keeps a summary of each day: how many people visited, and how many did each KPI.
Counting people from these summaries is accurate to about 1%. Needlr adds that tiny doubt to the result, so a verdict from summaries is a little more careful, never bolder. The check page says Older than 90 days: counted from daily summaries.
No history? Try Google Analytics
Connect Google Analytics on your project's Settings tab (the Google Analytics card). Its events can be the KPI for a release, and Google Analytics keeps their history.
- A release up to a year back can get a verdict right away, on every plan.
- In step 2, pick a KPI with the GA4 tag. Or, if step 3 warns about missing history, pick one of the Google Analytics events it offers.
- Google Analytics KPIs work only for Before vs after. They can't be used for a split test or an adoption check.
What step 3 may tell you
- This KPI is new, so Needlr has no history from before the release. It starts counting when you start, and needs 4 weeks before the release: pick a release date at least 4 weeks from now, or use a split test.
- Your plan measures releases up to a number of days back: 395 days, the full 13 months.
- Needlr needs 4 weeks of this KPI before the release. It says the earliest release date it can measure.
- Google Analytics has history from before the release. Count one of its events instead.
Common mistakes
- Making the KPI on the day you want to measure. It has no past. Make your main KPIs early.
- The wrong release date. Use the day it went live for everyone, not the day the work started.
- Two releases close together. Needlr warns you, but it can't tell them apart.
What the video says
Something shipped three months ago, and nobody measured it. An impact check can still give you a verdict, as long as Needlr was already counting that KPI before the release.
Needlr keeps a daily summary of every KPI it counts, for 13 months, on every plan. The oldest ones are measured from these summaries, accurate to about 1%.
Every account reaches back the full 13 months, so last spring's launch can still get a verdict. Did it move the needle? You can still find out.
The rule: Needlr needs four weeks of the KPI before the release. A KPI you created last month can't measure a release from last spring, because there's no before.
No history in Needlr? Google Analytics may have it. Connect GA4, pick one of its events, and reach back up to a year, with results right away.
The lesson: set up your key KPIs today, even with nothing to test. Every day Needlr counts is a day you can measure later.