Try it on your own data
Every report in this article is in the free plan. One snippet for web and apps, up to 10 sites.
Start freeSwitching analytics tools is technically trivial and organisationally hard. This is the checklist we use with teams moving from Google Analytics 4: eight checks in order, with the metric mapping, the parallel-run exit criteria, what to export before access lapses, and what to verify after the old tag is gone.
What you will learn
- An event and key-event inventory you can finish in an hour
- A metric mapping table, so the first week of comparisons does not turn into an argument
- A parallel run with exit criteria, a history export and a switch-date annotation
What changed · Nov 3, 2026
Rewritten as an 8-step GA4 migration checklist: definitions checked against Google docs, history export, monitors and an FAQ.
Before you migrate from Google Analytics1. Inventory events and conversions2. Map the metric definitions3. Run both tools in parallel4. Install cleanly, then map the custom events5. Rebuild funnels and dashboards6. Access, monitors and integrations7. Export the history you will want8. Cut over on a quiet dayFAQ
Before you migrate from Google Analytics
New section
When you migrate from Google Analytics 4 to another tool, the snippet takes a minute. What takes a month is the rest: the new numbers do not match the old ones, three dashboards still point at a property nobody updates, and a conversion that fired on desktop never fired on mobile. This checklist covers that month, in order. Each check makes the next one easier, and if the first seven are done, the cutover at the end is a ten-minute change.
1. Inventory events and conversions
Section updated
Export the list of events your property receives and mark the ones set as key events (conversions). Most properties have three kinds: automatic events nobody configured, custom events someone added for a campaign two years ago, and the five to ten that actually drive decisions. Only the last group needs to exist on day one. Write each down with the parameters a report depends on, and who reads it.
If your site has grown faster than your tracking plan, TapCub’s autocapture and event inbox proposes candidate events from real clicks and form submissions once the snippet is in. You accept, merge or ignore each one, which often surfaces the events the old plan missed. Name them by the rules in event naming conventions.
2. Map the metric definitions
Section updated
Most “the numbers are wrong” conversations in the first week are definition mismatches. Two tools count different things under similar names. In GA4, a session is engaged if it lasted longer than 10 seconds, had a key event or had two or more page views, and bounce rate is the share that were not engaged (Google Analytics Help). In TapCub, a bounce is a session with one pageview and no event. Those are different rules, so compare the direction by channel, not the absolute rate.
Write the mapping down before anyone compares a dashboard, and keep it in the same doc as the migration plan. The table below is the one we hand over.
| Google Analytics 4 term | TapCub term | What to compare |
|---|---|---|
| Users | Visitors (cookieless; recognized bots excluded) | Daily visitors and trend shape; monthly totals differ by design |
| Sessions | Sessions (30 min idle, 24 h cap) | Sessions per visitor, within a few percent |
| Engaged sessions / bounce rate | Bounce rate (one pageview, no event) | Direction by channel, not the absolute rate |
| Views | Pageviews | Top 10 pages by share; the ranking should match |
| Key events | Page goals, event goals, funnels | Counts once mapped; rates divide by visitors |
| Default channel group | Channels, plus your channel rules | Recreate custom groupings as channel rules |
3. Run both tools in parallel
New section
Install the TapCub snippet next to the existing tag and leave both running. The two libraries report to their own endpoints and do not read each other’s data. Run for at least one full week; two to four weeks if you report monthly, so you see the reporting cycle once in both tools.
Compare shape, not totals: do the daily trend lines move together, do the same pages lead, do the same sources rank? Compare daily visitors, where both tools agree on what a person is, and write down the ratio rather than the difference. The remaining gap is explained by why analytics tools show different numbers: bots, consent, cookies and session rules.
Set exit criteria on day one, for example “daily visitors within 15% once bots are excluded, top ten pages match in order, all five conversions firing”. A parallel run without exit criteria never ends. The sample board below shows what that looks like midway.
Parallel run · exit criteria
Day 14 of 28| Criterion | Target | Current | Status |
|---|---|---|---|
| Daily visitors, humans only | within 15% | 11% | Met |
| Top 10 pages match in order | 10 of 10 | 9 of 10 | Close |
| Top 5 sources match | 5 of 5 | 5 of 5 | Met |
| Conversions firing | 5 of 5 | 4 of 5 | purchase missing on mobile |
| Alerts re-created | 6 | 6 | Met |
Sample migration board. The missing purchase event on mobile was a tag-manager trigger scoped to desktop.
4. Install cleanly, then map the custom events
Section updated
Put the snippet in the head of every template, or use the plugin for your site builder; the installation guide covers tag managers and single-page apps. Then map each event from the inventory to a track() call with the same name and parameters, so year-over-year notes stay readable. Map each key event to a page goal, an event goal or a funnel; for purchase amounts you need exact, send a server-side event.
UTM parameters need no mapping. Ad click IDs from 14 platforms, including gclid and msclkid, are read automatically. If your team relies on a custom channel grouping, recreate it as channel rules, which take precedence over the defaults.
// keep the old names so year-over-year comparisons stay readable
tapcub.track('begin_checkout', { value: 89, currency: 'USD' });
tapcub.track('purchase', { value: 89, currency: 'USD', order_id: 'A-10422' });
tapcub.track('generate_lead', { form: 'contact' });
// identify logged-in users for retention and profiles
tapcub.identify('user_8f21', { plan: 'pro' });5. Rebuild funnels and dashboards
Section updated
List every dashboard, scheduled report and funnel that points at the old property, and who owns each. Rebuild the ones people actually open; retire the rest. This is usually when teams find that half their reports have not been viewed in a year. Define funnels on visitors with a conversion window, following the conversion funnel checklist.
Build for the people who read them, not as copies of the old ones. A weekly leadership view needs five numbers and a trend; a marketing view needs channels, campaigns and paid clicks; a product view needs the activation funnel and a retention matrix. One thing to state in every report title: TapCub goal conversion rates divide by visitors, so a GA4 report that divided by sessions will show a different percentage for the same count.
6. Access, monitors and integrations
New section
Invite the team and set roles before the cutover, not after someone cannot find a report. Re-create the alerts you rely on as monitors: traffic drops, error spikes, a collection gap, or a saved conversion metric falling below a threshold. A dry run shows how often each rule would have fired over the last 30 days, which helps you tune it before the first real alert. Reconnect the integrations that feed or read your data.
7. Export the history you will want
New section
Historical totals should not be loaded into the new tool. They were counted with GA4’s definitions, so the trend line would jump on the switch date for no real reason. Keep read access to the GA4 property instead, and export what you will quote later as CSV: monthly users and key events for the last two or three years, and top pages and sources by quarter.
Two GA4 details matter here. The data retention setting (2 or 14 months for standard properties) limits explorations and funnel reports, not the standard aggregated reports (Google Analytics Help), so check which ones your team uses before access lapses. And add an annotation on the switch date in the new tool, so a step in any chart is read as a tool change, not a traffic change. TapCub keeps raw data for 6 months on Free, 12 on Pro and 24 or more on Max and Ultra; create raw exports for a longer archive.
8. Cut over on a quiet day
Section updated
Pick a low-traffic day, make the new dashboards the links people open, and confirm that the day’s numbers look like the days before. Keep the old tag for one more month as a fallback, then remove it from the template, not only from the tag manager if it was ever hard-coded, and update the cookie list in your consent banner.
Expect one or two surprises in the following weeks anyway: a report somebody built privately, a landing-page tool still firing the old tag, a monthly export nobody knew about. Keep the migration doc open for a month and add each one to the checklist for next time.
FAQ
New section
How long should GA4 and the new tool run in parallel?
At least one full week, so you see every weekday in both tools. If you report monthly, run two to four weeks and end the run when your exit criteria are met, not on a fixed date.
Can I import my GA4 historical data?
Not one for one. GA4 history uses its own user, session and channel definitions, so loading totals into a tool with different rules creates a false jump on the switch date. Keep read access to GA4 and export the reports you quote as CSV.
Why does the visitor count differ from GA4 during the parallel run?
The tools count bots, consent, cookies and sessions differently. Compare daily visitors and the trend shape week over week; the totals will not match, and they do not need to.
Do the two tags slow the page down or interfere?
They do not read each other’s data. The cost is the extra script; the TapCub snippet is about 2 KB gzipped and loads with defer, so it does not block rendering.
What should I check after removing the GA4 tag?
That the day’s numbers look like the previous days, all key events still fire on desktop and mobile, monitors are active, and the consent banner no longer lists GA4 cookies.
See also: Data migration project: a five-phase plan with reconciliation checks.
See also: What is UA? Universal Analytics, user acquisition and user agent.
Update history1
- Major updateRewritten as an 8-step GA4 migration checklist: definitions checked against Google docs, history export, monitors and an FAQ.
- First published
Was this article helpful?
We use your answers to fix and expand this page.
Couldn’t send. Please try again in a minute.
Thanks for the feedback. We’ll use it to improve this page.You’ve already sent feedback on this page. Thanks.

Solutions Engineer
Covers SDKs, API integrations, and data implementation, with an emphasis on validation, system interoperability, and troubleshooting.

Data Engineer
Writes about event design, data quality, and collection workflows, with verifiable implementation methods for reliable analytics.



