How to keep historical data when switching analytics tools
Decide what is worth keeping
The most common mistake is trying to keep everything. For a small site, a few aggregated exports usually answer the questions you will still ask in a year.
Before you cancel the old tool, list the reports you actually use:
- visitors and pageviews by day, week, or month
- top pages by visits and unique visitors
- traffic sources and campaigns
- goal completions or signups
- conversion rates by landing page
- referrer or social network breakdowns
If you only look at monthly traffic and top pages, export those. If you run seasonal campaigns, keep campaign and landing page data at least back to the same period last year.
Export before you remove the tag
Export everything while you still have access. Do not assume you can log in after cancellation or that the vendor will keep your data.
Most tools let you export reports from the dashboard or through a reporting API. Check your tool's own documentation for the exact steps. Export at the most detailed level you might need. A daily CSV with one row per page and source is more useful later than a screenshot of a chart.
For each export, record:
- the tool and account
- the site or property name
- the date exported
- the timezone
- what a visit or conversion meant in that tool
- any filters or segments applied
A short README file next to the CSV prevents future confusion.
Run both tools in parallel for a few weeks
Before you remove the old script, add the new one and let both run for at least two to four weeks. This overlap period gives you a side-by-side view.
Expect differences. A cookieless tool and a cookie-based tool may count visitors differently. Bot filtering, session timeouts, ad blockers, consent banners, and daily anonymous ids can all change the numbers. The goal is not to make the tools match exactly. The goal is to learn roughly how the new numbers compare to the old ones.
Record the exact date the new tool started collecting data. When you later see a drop or spike at that point, you will know it is a measurement change, not a traffic change.
Store the old data in a neutral format
Do not keep old analytics only as screenshots or proprietary exports. Convert what you can into CSV or JSON Lines.
A simple CSV shape works for most small sites:
date,path,source,visitors,conversions 2026-01-01,/,google,120,3 2026-01-01,/pricing,direct,40,1
Use UTC or clearly label the timezone. Avoid merged cells, pivot tables, and color-coded spreadsheets as your only archive. Clean tables are easier to combine with new data later.
Keep the old account read-only for a transition window
If the old tool allows read-only access after you stop collecting data, keep it for at least a few weeks or months. You may need to re-export a report you forgot, or check a definition.
Check the tool's terms and your own legal position before you delete anything. Some tools reduce access after cancellation, so finish your exports first.
Combine old and new data for reporting
Do not try to force old numbers and new numbers into one continuous line without context. Instead, combine them in a report with a column that shows which tool produced each row.
For a simple trend, use old data through the day before the switch and new data from the switch date onward. For year-over-year comparisons, you need the old baseline. If your new tool starts fresh today, you will not have last year's data there, so keep the old CSV archive available.
If you use a coding agent, ask it to merge the old CSV files with the new reports and produce a trend chart or weekly summary. Some analytics tools can expose their data to an agent over MCP. measuremy.site, for example, has an MCP server that lets Claude Code, Cursor or Codex read the analytics. That can make the combined reporting less manual.
Preserve campaign history
Historical traffic source data is easy to lose in a migration. Export source, medium, campaign, and landing page reports before switching. If you use UTM-tagged links, keep a list of the tags you have used and what they mean.
If you created tracked links in the old tool, export their destinations and slug or short-code values if you may need to recreate or redirect them. Do not rely on the old tool's link redirects working forever after cancellation.
Document the switch
Add an annotation in the new tool if it supports annotations. If not, keep a simple log file:
- old tool stopped: date
- new tool started: date
- overlap period: date range
- main metric differences observed
This stops future-you from misreading the switch as a traffic crash or spike. It also helps if someone else starts looking at the dashboard later.
Set a retention plan for the new data
Once the old data is archived, decide how long you will keep the new data. If you want year-over-year comparisons, keep at least 13 months of daily or monthly aggregates. If your privacy policy or legal situation requires shorter retention, follow that and keep only aggregated data that is still useful for trends.
The old archive can stay in cold storage. It does not need to be in the new dashboard to be useful.
A practical switch checklist
- Export all reports you currently use
- Save the most detailed aggregate data you can
- Note definitions, timezone, and export date
- Run old and new tools in parallel for a few weeks
- Record the switch date
- Keep the old account read-only until the transition is done
- Archive the old data in CSV or JSON Lines
- Document why the numbers changed
Switching tools is easier when you treat the old data as an archive, not as something to import perfectly into the new system. Keep the raw material, preserve the definitions, and run both tools long enough to understand the relationship between their numbers.
See it on your own site
measuremy.site is cookieless analytics your coding agent can read. 30-day free trial, no card. Plans from $9/month.
Start free trial See the live demoAdd measuremy.site analytics to this project: connect the measuremy.site MCP server at https://a.measuremy.site/mcp for all my projects (claude mcp add --scope user; it signs in with OAuth, no key, and one sign-in covers every site), then tell me how to sign in. Details: measuremy.site/llms.txt
Free tools
- Analytics tag checker: Which analytics and pixels a site runs, plus duplicate and broken tags.
- Free UTM link builder: Add utm_source, medium and campaign to any link.
- Website traffic checker: A site's popularity rank and estimated monthly visits.
More from the blog
- How to calculate your website conversion rate, and what to compare it to
- Is my A/B test result significant? A plain-English guide with a worked example
- What is bounce rate, how is it calculated, and is yours a problem?
- How to add a visitor counter to your website or GitHub README
- How much traffic does a website get? How to estimate any site's visits
- How to track QR code scans: count them, and see which poster worked
- How to see how many people clicked your link (without a paid shortener)
- How to check if your Google Analytics 4 tag is installed (and only counting once)
- Why GA4 shows your campaign traffic as Unassigned (and how to fix the links)
- Does your redirect keep your UTM tags? How to check before a campaign goes out
- How I built a self-hosted analytics MCP server on Cloudflare Workers and D1
- UTM tracking explained: what the parameters mean and how to use them
- Conversion funnel analysis for small sites: a practical guide
- Sudden drop in website traffic? A checklist for finding the cause
- What is an MCP server? A plain-English guide for people who run websites
- Vanity Metrics vs Useful Metrics: What to Track
- How to Compare Cookieless Website Tracking Tools
- Why Your Analytics Numbers Differ Between Tools
- When GA4 Is Too Much for a Small Website
- How to give an AI agent read-only access to your analytics
- Why developers are connecting agents to production data
- How to recover from a sudden traffic drop
- Website analytics for indie hackers: what to watch and what to ignore
- Direct Traffic Explained: Why So Much Is Unknown
- Why a high bounce rate is not always bad
- High bounce rate: 8 causes and practical fixes
- How to Set Up Traffic Alerts for Sudden Drops
- How to Track Traffic From a Podcast
- How to measure the return on a paid ad campaign
- How to use analytics to decide what to write next
- How the Model Context Protocol Works
- Track File Downloads Without Guesswork
- Adding analytics to a Next.js app without slowing it down
- Core Web Vitals and Traffic: Is There a Link?
- Analytics After Cookies: What Actually Changes
- GDPR and website analytics: a practical guide for small sites
- How to Compare Website Analytics Tools by Privacy
- GA4 Events Explained for Beginners
- How to measure a newsletter send with UTM links
- How to track Hacker News traffic
- Adding Analytics to a Framer Site
- How to ask your coding agent about your analytics
- What Is Bounce Rate and How Is It Calculated?
- A practical website analytics retention policy
- Map your site's funnel from first visit to signup
- What Is a Good Bounce Rate?
- How to automate a weekly traffic report
- A Practical Privacy-Friendly Analytics Checklist
- Adding an MCP server to Claude Code
- How long should you keep analytics data?
- How to Build a Small MCP Server That Actually Helps
- How to Find the Biggest Drop Off in a Funnel
- What Is the Urchin Tracking Module?
- What Microsoft MCP servers actually do
- UTM parameters for email newsletters
- Track Button Clicks Without Google Tag Manager
- How to track the results of a social media post without cookies
- How to track traffic from a GitHub README
- How to measure a product launch without guessing