Why a site migration puts organic traffic at risk
Sooner or later every website gets rebuilt. The shop moves to a new platform, a redesign changes the menus, a rebrand renames the sections, or someone simply decides that the addresses of the pages should be tidier. For visitors this is a visual change. For search engines it is something more serious: the addresses they have stored for years, and the rankings attached to them, suddenly lead nowhere. Done well, a migration costs a short dip in organic traffic that recovers within weeks. Done badly, it can remove a large share of the traffic permanently.
The reason is simple. A search engine does not know that /guides/boot-fit and /guides/footwear/fit-guide are the same page unless you tell it. It has to be told with a redirect from the old address to the new one, and the redirect has to be in place on the day the old address stops working. Every old URL without a sensible successor is a small leak. Add up the leaks of the busiest pages and you have the migration risk in numbers.
This free lesson teaches the first and most important document of any migration, the URL map, and the first number you calculate from it: redirect coverage, measured in two ways. The example is Halvergate Outfitters, a fictional online shop for outdoor gear. All URLs, sessions and companies in this course are invented for learning, and benchmark values are illustrative.
The core concept: one row per old URL
A URL map is a table with one row for every old URL that matters, and at least three columns: the old URL, the organic sessions it received in a recent period, and the new URL it should redirect to. If there is no planned target yet, the cell says none. The old URLs come from several sources, and the union of them is your inventory: the analytics landing-page report, the search console's page report, the XML sitemap, a crawl of the live site and a list of URLs with backlinks. A page that earns no sessions today can still carry links, so do not rely on a single source.
Next to each URL you record the section (tents, boots, guides) so that problems can be grouped later. The sessions column is what turns the list into a decision tool: it tells you which rows deserve the most care. A 90-day window smooths out single-week spikes, and using organic sessions only keeps paid and email traffic from distorting the picture.
Two labels follow from the target column. A URL is Mapped when it has a target and a Gap when it has none. A third, more detailed layer (is the target really the closest equivalent page? does it answer the same search intent?) comes in later lessons. For now the question is only: does every important old URL have somewhere to go?
A fully worked example: Halvergate Outfitters
Here are the twelve old URLs of Halvergate Outfitters with their sessions over the last 90 days and the redirect target that the project team has planned so far. Three rows say none. The last column is the URL's share of all organic sessions in scope, which is 32,580.
| Old URL | Section | Sessions (90 days) | New URL | Status | Share |
|---|---|---|---|---|---|
/tents/alpine-2p | Tents | 3,820 | /camping/tents/alpine-2p | Mapped | 11.7 % |
/tents/family-4p | Tents | 2,110 | /camping/tents/family-4p | Mapped | 6.5 % |
/boots/trail-low | Boots | 3,160 | /footwear/trail-low | Mapped | 9.7 % |
/boots/winter-high | Boots | 2,240 | /footwear/winter-high | Mapped | 6.9 % |
/jackets/shell-light | Jackets | 2,650 | /clothing/shell-light | Mapped | 8.1 % |
/jackets/down-parka | Jackets | 1,580 | /clothing/down-parka | Mapped | 4.8 % |
/packs/day-25l | Packs | 1,120 | /packs/daypacks/25l | Mapped | 3.4 % |
/packs/trek-60l | Packs | 3,480 | none | Gap | 10.7 % |
/guides/choose-a-tent | Guides | 5,240 | /guides/tents/choose | Mapped | 16.1 % |
/guides/boot-fit | Guides | 6,130 | none | Gap | 18.8 % |
/sale/summer-clearance | Sale | 740 | none | Gap | 2.3 % |
/about/our-story | Brand | 310 | /about | Mapped | 1.0 % |
| Total | 32,580 | 100.0 % |
Read the table before calculating anything. The busiest URL is the boot fit guide with 6,130 sessions, and it has no target. The second busiest, the tent-buying guide with 5,240, is mapped. The product page for the 60-litre trekking pack has 3,480 sessions and also no target. The only other gap is the seasonal clearance page with 740 sessions. Three gaps out of twelve URLs sounds manageable, but the next steps show how much traffic those three represent.
The formulas, step by step
Everything in the lab uses a handful of spreadsheet functions. Written for row 4 and copied down to row 15, the formulas are:
- Status (E4):
=IF(D4="none","Gap","Mapped"). The text in column D is compared with the word none. Note that the cell is not empty, it contains a word, so a test likeD4=""would label everything Mapped. - Share of sessions (F4):
=C4/SUM($C$4:$C$15). The dollar signs lock the range, so every row divides by the same total. Without them the range slides down and the shares are wrong. - Coverage by URL count (I5):
=COUNTIF($E$4:$E$15,"Mapped")/I4, where I4 holds=COUNTA(A4:A15), the number of URLs in scope. - Coverage by sessions (I6):
=SUMIFS($C$4:$C$15,$E$4:$E$15,"Mapped")/SUM($C$4:$C$15). SUMIFS adds the sessions of only those rows whose status is Mapped.
Take the boot fit guide as a check: its share is 6,130 ÷ 32,580 = 18.8 %, the largest share of any URL. The three gap rows add up to 10,350 sessions, which is 31.8 % of the total. Keep that number in mind.
Coverage by URL count versus coverage by sessions
There are two natural ways to ask how well the map covers the site. Coverage by URL count divides the mapped URLs by all URLs: 9 ÷ 12 = 75.0 %. It is easy to explain and easy to track in a project plan. But it treats every URL as equal, and in search that is never true. A handful of pages usually earn most of the traffic, and the long tail of pages earns very little.
Coverage by sessions weights each URL by what it earns: the sessions of the mapped URLs ÷ all sessions = 22,230 ÷ 32,580 = 68.2 %. The gap between the two numbers is the message. At Halvergate Outfitters the team could report that three quarters of the URLs are covered, and it would be true, while almost a third of the organic traffic points at addresses with no plan.
The direction of the difference matters too. Here sessions coverage is lower than URL coverage, because the gaps are concentrated in busy pages (the two big gaps are the first and the fourth busiest URLs of the twelve). On another site the long tail could be the problem and the two numbers would be closer or reversed. You do not know until you calculate both, so make the session-weighted number the headline and the URL count the supporting detail.
How to read the result and what to do with it
A coverage figure is not a verdict, it is a work list. Sort the gap rows by sessions and work from the top. For each gap the questions are the same: is there a page on the new site that answers the same need? If yes, point the redirect there. If the content was dropped on purpose, choose the closest relevant page (usually the parent category or a related guide), and accept that some value will be lost. Redirecting everything to the home page is the classic shortcut, and it is a poor one: it sends visitors and signals to a page that does not answer what they were looking for.
- /guides/boot-fit (6,130 sessions, 18.8 %): the highest priority. If the new site has a footwear fit guide, redirect there. If not, the guide should be rebuilt before launch.
- /packs/trek-60l (3,480 sessions, 10.7 %): a product page. Redirect to the successor product, or to the trekking-pack category if the product is discontinued.
- /sale/summer-clearance (740 sessions, 2.3 %): a seasonal page. Redirect to the current sale hub; its traffic fades anyway.
- All mapped rows: check that each target returns a normal page and that the redirect is a single hop (lesson 4 covers chains and dead ends).
Step by step in the spreadsheet
- Paste the old URLs into column A, the section into B and the 90-day organic sessions into C. Make sure C holds numbers, not text: they should be right-aligned.
- Type the planned target in column D. Where no target exists, type the word none (always the same word, in the same spelling, so formulas can find it).
- Click E4 and type
=IF(D4="none","Gap","Mapped"). Copy it down to E15 with the fill handle or Ctrl + D. - Click F4 and type
=C4/SUM($C$4:$C$15). Press F4 (Mac: Cmd + T) after selecting the range to add the dollar signs. Format the column as a percentage and copy it down. - In I4 type
=COUNTA(A4:A15). In I5 type=COUNTIF($E$4:$E$15,"Mapped")/I4. In I6 type=SUMIFS($C$4:$C$15,$E$4:$E$15,"Mapped")/SUM($C$4:$C$15). Format both as percentages. - Cross-check: the shares in column F must add up to 100 %.
=SUM(F4:F15)tells you at once whether a range was left unlocked. - Select A3:C15, choose Insert › Chart and draw the sessions per URL as columns or bars. The gaps become visible when you sort the table by sessions first.
Common mistakes and how to spot them
- Testing for an empty cell. If the target column says none,
=IF(D4="","Gap","Mapped")finds no empty cell and labels every URL Mapped. The coverage then shows 100 %, which is too good to be true. - Unlocked totals.
=C4/SUM(C4:C15)copied down becomesSUM(C5:C16)in the next row. The shares drift and do not add up to 100 %. Lock the range with dollar signs. - Counting URLs and calling it traffic. A coverage of 75 % by URL count says nothing about how much traffic is protected. Always give the session-weighted figure next to it.
- Using one metric only. Sessions are not the only asset an old URL carries: backlinks, conversions and the keywords it ranks for matter as well. Lessons 5 to 7 add those views.
- Mapping to the home page. A redirect to an unrelated page is rarely better than none. Aim for the most relevant equivalent.
- Stale sessions. A URL map built from last year's data misses pages that became popular since. Refresh the export a week before launch and re-check the gaps.
Beyond coverage: what else a good URL map holds
Coverage is the first column of a bigger document. As the project matures the same table gains more fields, each of which becomes a lab in this course. You will recognise them when you reach them:
- Redirect type and hops (lesson 4): permanent or temporary, one hop or a chain, and whether the final page loads.
- Backlink value (lesson 5): referring domains and their strength, because links are an asset that follows the redirect.
- Revenue per session (lesson 6): to rank the work by money at risk instead of by page count.
- Keyword positions (lesson 7): the queries each URL ranks for and the clicks they bring.
- Quality checks on the new URL (lesson 8): indexable, canonical to itself, no broken links, ready for the launch gate.
Quick reference: functions and shortcuts in this lesson
| Function / shortcut | What it does here | Example |
|---|---|---|
IF | Labels a URL Mapped or Gap | =IF(D4="none","Gap","Mapped") |
SUM | Total sessions, the denominator of the shares | =SUM($C$4:$C$15) |
COUNTA | Counts the URLs in scope | =COUNTA(A4:A15) |
COUNTIF | Counts the Mapped rows | =COUNTIF($E$4:$E$15,"Mapped") |
SUMIFS | Adds the sessions of the Mapped rows | =SUMIFS($C$4:$C$15,$E$4:$E$15,"Mapped") |
| Absolute reference | Locks a range so that copying does not move it | $C$4:$C$15 (F4 key) |
| Fill down | Copies a formula to the rows below | Ctrl + D |
Real-world uses
- Migration project plan. The URL map is the central deliverable. Share the coverage figures weekly with developers and stakeholders, and block the launch date until the standard is met.
- Redesign without a platform change. Even a facelift that renames categories changes URLs. The same table protects the traffic of a small redesign.
- Merging two sites. When two shops or two blogs are combined, the URL map shows which sections overlap and where targets conflict.
- Agency and client communication. A single number, coverage by sessions, is far easier to discuss with a non-technical client than a crawl export.
- Post-launch audits. Re-run the sheet after launch with the new crawl: any old URL that still has no working target is a leak that can be closed this week.
What you will build in the rest of this course
Lesson 1 is free. Lessons 2 to 20 make up the paid course. Each has its own live lab with a different data set from the same fictional company and builds on this map, all the way to a migration scorecard.
Traffic at Risk: Expected Loss per Redirect Type, Sessions at Risk and the Share of Traffic You Could Lose. You attach an expected loss to every redirect type and size the traffic at risk before anything is launched.
Pre-Launch Benchmarks, Redirect Audits and Link Equity. You freeze a baseline and an expected line from last year's seasonality, audit chains, temporary redirects and dead ends, and value the backlinks that depend on each redirect.
Priorities, Keywords and the Launch Gate. You rank the work by revenue at risk with a cumulative 80 % list, translate keyword positions into expected clicks and build a go or no-go gate from a staging crawl.
Launch-Day Monitoring, Indexing and Recovery. You track hourly errors, the handover of old to new URLs in the index, the weekly gap against the benchmark and a recovery curve with a straight-line forecast.
Diagnosis, Revenue and Rollback. You test whether a drop is real, find the worst segment, separate lost traffic from lost conversion and set rollback thresholds that turn the metrics into a decision.
Waves, Scenarios, Backlog and Scorecard. You plan a phased launch with WORKDAY, build a what-if grid with a heat map, rank post-launch fixes by value per hour and finish with a migration scorecard.
Summary
A migration starts with a URL map: one row per old URL, its sessions and its planned target. =IF(D4="none","Gap","Mapped") labels each row, =C4/SUM($C$4:$C$15) gives each URL's share of the traffic, and COUNTIF and SUMIFS turn the labels into two coverage figures. At Halvergate Outfitters 9 of 12 URLs (75 %) are mapped, but only 68.2 % of the sessions, because the three gaps include the busiest URL of all. Always report the session-weighted coverage, and close the biggest gaps first.
Now scroll up to the lab and build the sheet yourself. It takes about fifteen minutes, the Coach checks each step, and you will end up with a coverage check you can apply to your own URL export.
Frequently asked questions
What is a URL map in a site migration?
A URL map is a table with one row per old URL, showing the traffic it receives and the new URL it should redirect to. It is the central planning document of a migration, because every old URL without a sensible target is a place where search traffic can be lost.
What is redirect coverage?
Redirect coverage is the share of old URLs that have a planned redirect target. It can be measured by URL count (mapped URLs divided by all URLs) or weighted by sessions (sessions of mapped URLs divided by all sessions). The session-weighted figure shows how much traffic is protected.
Why is coverage by sessions lower than coverage by URL count in this example?
Because the URLs without a target include the busiest URL of all (the boot fit guide) and a product page with 3,480 sessions. Counting URLs treats them like any small page, while weighting by sessions shows that the three gaps carry almost a third of the organic traffic.
How do I calculate the share of sessions for each URL in Excel?
Divide the URL's sessions by the total with an absolute range: =C4/SUM($C$4:$C$15), then copy the formula down. The dollar signs keep the total range fixed so that every row divides by the same number.
Should I redirect every unmapped URL to the home page?
Usually not. A redirect should point to the closest relevant page, such as the successor product, the parent category or a related guide. Redirecting unrelated pages to the home page rarely answers what the visitor was looking for and is no substitute for proper mapping.
Where do I get the list of old URLs and their sessions?
Combine several sources: the landing-page report of your analytics tool, the page report of the search console, the XML sitemap, a crawl of the live site and a list of URLs that have backlinks. Use organic sessions over a window such as 90 days and refresh the export shortly before launch.
Keep going: SEO for Site Migrations & Redesigns: URL Mapping, Redirect Coverage, Traffic at Risk, Benchmarks, Recovery Curves & Rollback Thresholds
This free lab is lesson 1 of 20. Unlock the other 19 hands-on labs, the final exam and your certificate for a one-time $9.90.
- Traffic at Risk: Expected Loss per Redirect Type, Sessions at Risk and the Share of Traffic You Could Lose
- Pre-Launch Benchmarks: Baseline Level, Year-on-Year Factor and Expected Sessions for the Weeks After Launch
- Redirect Audit: Chains, Temporary Redirects and Dead Ends, Counted Per Issue and Weighted by Sessions
- Link Equity at Risk: Link Value Index, Value Kept per Redirect Type and the Share of Backlink Strength You Could Lose
- Priority Ranking: Revenue at Risk per Page Group, RANK.EQ, Cumulative Share and the 80 % List
- Keyword Rankings Before and After: Click-Through Curve Lookup, Expected Clicks and the Change in Organic Visibility
- Staging QA Gate: Issue Rate per Template, Blocker Status and a Go or No-Go Launch Decision
- Launch-Day Error Monitoring: Hourly Error Rate, Amber and Red Alerts, Cumulative 404s and the First Red Hour
- Index Coverage After Launch: Old vs New URLs Indexed, Share of Baseline, Indexing Pace with SLOPE and the Day the Target Is Reached
- Post-Launch Gap: Actual vs Expected Sessions, Percent of Benchmark, Cumulative Shortfall and Revenue Lost
- Recovery Curve: Weekly Gain, Pace with SLOPE, a FORECAST.LINEAR Projection and the First Week at Target
- Is the Drop Real? Pre-Launch Noise Band, STDEV.S, z-Scores and a Real-Drop Verdict per Week
- Segment Diagnosis: Change per Segment, Percentage Drop, Share of the Net Change and the Worst Segment with INDEX-MATCH
- Revenue Impact: Expected vs Actual Organic Revenue, the Eight-Week Gap and the Share Caused by Lost Traffic
- Rollback Thresholds: Direction-Aware Red and Amber Status, Risk Score and a Roll Back, Hold or Continue Decision
- Phased Migration Waves: Sessions at Risk per Wave, WORKDAY Launch Dates, Cumulative Share Migrated and Peak Exposure
- Scenario Grid: Revenue Lost for Every Initial Drop and Recovery Time, a Heat Map and the Longest Affordable Recovery
- Fix Backlog: Value per Hour, RANK.EQ, Cumulative Hours Against Sprint Capacity and the Value of the Sprint
- Final Project: Migration Scorecard with Weighted Readiness, Percent of Benchmark, Revenue Lost, a Decision and Charts