How to Compare LSA Performance Before and After Migration

A fair LSA performance comparison uses three periods: before migration, transition, and after migration. Save the last 28 complete pre-migration days. Keep migration day through Day 14 in a separate transition record. Start the 28-day post period after that transition.

Google's LSA migration announcement says many ads may begin running at once, while migration and campaign performance can take up to two weeks to return to stable levels. Use that period to verify transfer accuracy and lead handling. Preserve its data, but keep it out of the durable 28-day comparison.

Match the weekday mix, service categories, locations, and metric definitions across the before and after periods. Add a prior-year period when a seasonal business has dependable history. Record weather, staffing, schedule, promotions, territory, and capacity changes with dates.

Save the old campaign report before the migration deadline. Google's Local Services Ads report guide documents charged leads, lead category, Category or Direct Business Search attribution, call, message, or booking type, charged amount, credits, and booked-job details entered by the advertiser.

Name the export with the Customer ID, location, start date, end date, and export date. Open the file and confirm the account, period, fields, and totals. The Where old LSA reports go after migration explains the separate destinations for historical performance, past lead history, and new Google Ads reporting.

Export business outcomes for the same dates from the approved CRM, intake, booking, dispatch, or accounting system. Keep the raw Google report and the raw business record unchanged. Build the comparison in a working file that cites both sources.

Use one dictionary across the three periods. Define the numerator, denominator, source, owner, and date basis for each measure. Keep Google campaign measures, lead operations, and business outcomes in separate layers.

Google campaign data answers how much the account spent and how many charged leads it recorded. Lead operations add call, message, booking, charged status, feedback, answer, and response information. Business systems add qualification, booked jobs, completed work, revenue, and lost-reason data.

A missing denominator should produce Not available. Entering zero implies a measured result. Keep revenue on a stated lead-date, booking-date, or completion-date basis so the two periods use the same rule.

Label migration day as Day 0. Use Days 0 through 14 to compare the expected campaign with the found campaign, confirm Lead Manager access, test call and message routing, and review verification status. Record any pause and the time it ended.

Open Campaigns > Settings and compare services, areas, schedule, daily budget, campaign Target CPA, and status with the saved inventory. Open Goals > Conversions > Leads and verify lead rows, charged status, lead types, messages, and permitted exports. Open Assets for photos and callouts.

Place Google Ads Change history beside the dates. An edit to budget, bidding, targeting, or status can explain part of the movement. Keep the user, timestamp, old value, new value, and reason with the performance worksheet.

Open the migrated campaign table and add Conversions, Cost, and Cost / conv. Select the date control, turn on Compare, and choose Previous period, Previous year, or a Custom range. Apply the dates, then expand the column comparison to see the current value, comparison value, numerical change, and percentage change.

Google's date-range comparison guide confirms those comparison controls. Google also notes that its explanations feature requires periods within the past 90 days that have equal length and touch each other. Your own migration worksheet can compare other approved periods, but it should state the dates and method.

Use Day-of-week segmentation when the business changes weekend or after-hours coverage. Keep full-day boundaries. A partial migration day can produce a false daily rate when it sits beside complete days.

Compare equal 28-day periods when the account has enough history. Twenty-eight days preserve the same weekday count on both sides. For a seasonal trade, add the same period from the prior year or another period with similar demand and capacity.

Record known differences instead of forcing a perfect match. Weather, holidays, staffing, office closures, promotions, prices, service areas, and technician capacity can affect lead volume and booked outcomes. Add each event with a date and source.

Small samples can move sharply when one or two jobs change. Report the counts beside the rates. Reserve statistical significance language for a defined method with enough observations.

Google converts the old weekly average into a daily average by dividing by seven. It sets the monthly spending limit at the daily average multiplied by 30.4. Compare actual spend for the same number of complete days instead of comparing the weekly and daily labels.

Calculate spend per complete day and leads per complete day when period lengths differ. Use those rates for diagnostic context, then return to equal periods for the main decision. Keep a note for high-demand days and any budget edit.

Review Cost beside the saved weekly budget, expected daily average, monthly limit, and Change history. A pacing difference can affect spend without proving that migration changed lead quality.

Add answer rate, message response, qualification, booking, completed jobs, revenue, and lost reasons to the charged-lead comparison. Use the same phone system, intake definition, booking system, and revenue basis across both periods.

A stable Cost / conv. can hide a booking problem. A higher Cost / conv. can still produce acceptable economics when qualification, booking, or job value improves. Review the counts and denominators before changing the campaign.

Use the fictional values below to follow the calculation method. They are not an LSA benchmark and do not prove that migration caused the movement. In the example, a small platform CPL change sits beside a larger downstream booking-cost change.

Use the downloadable template as a blank operating file. Enter one row for each account, location, and service category that shares stable definitions across the three periods. The formula cells calculate cost per charged lead, qualified rate, booking rate, and cost per booked job.

Cost per charged lead equals spend divided by charged leads. Qualified rate equals qualified leads divided by charged leads. Booking rate equals booked jobs divided by charged leads. Cost per booked job equals spend divided by booked jobs. The template returns Not available when a denominator has no value.

Keep customer contact details outside the worksheet. Use account IDs, categories, locations, and approved aggregate outcomes. Record source files and operational notes so another authorized reviewer can reproduce the comparison.

A setting mismatch requires transfer correction before a performance verdict. Compare services, areas, schedule, routing, status, budget, and Target CPA with the saved inventory. Record the corrected field and review date.

A volume change with stable quality points the reviewer toward visibility, spend, demand, service mix, and capacity. A quality change with stable charged-lead volume points toward lead type, query fit, call handling, message response, qualification, and booking.

Movement in both volume and quality needs a staged review. Start with material transfer and routing errors, then inspect Change history and operating context. Keep optional campaign edits in a controlled sequence so the next comparison has a readable cause record.

Source review completed July 30, 2026. Migration timing, report loss, new campaign metrics, budget conversion, Lead Manager, and stabilization guidance come from Google's LSA migration announcement. Historical report fields come from Google's Local Services Ads report guide. Google Ads comparison controls come from Google's date-range comparison guide.

The three-window method, 28-day cadence, metric dictionary, worksheet, formulas, and outcome framework are PrimeLSA operating recommendations. The numerical example is fictional and demonstrates calculation method. It is not a benchmark or evidence of a migration effect.