How-to guide

Line up ad spend against leads for every site in one ledger

You split one paid budget across many sites, and comparing ad spend vs leads by location is where it goes wrong: Google Ads, Local Services Ads and GA4 each count leads their own way. Build one ledger that counts every lead once and shows where the next dollar belongs.

Continue with Google

Free for 2 sites · No credit card

Or have your agent set it up

Set up DuckDive for my website. Follow https://analytics.duckdive.ai/agents.md

Updated 16 min read

The short answer

Give each site one row. For the same fixed window, pull Google Ads spend and conversions, Local Services Ads spend and charged leads, and GA4 conversions. Count paid search leads only under Ads. Divide spend by leads per channel to decide where money moves, and keep blended cost per lead for reporting the site as a whole.

  • Channel cost per lead decides budget moves; blended cost per lead reports results.
  • GA4 key events from ad clicks are already in the Ads column, so subtract them from GA4.
  • Move money on what the most recent spend bought, not on the average.

Build the spend-to-lead ledger

A paid budget split across sites gets argued over in three places: Google Ads, the Local Services Ads account and GA4. Each counts leads differently, so the cheapest site depends on which screen is open. A ledger ends that: one row per site, ten number columns, all for one window such as the last 30 days. Seven columns are read from Google; three are arithmetic.

Ads spend
The Google Ads account's cost for the window, from get_google_ads_summary. Every campaign in that account is included.
Ads conversions
Google Ads conversions as DuckDive reports them, from the same call. In Google Ads' own Conversions column these are primary actions filed by click date.
Ads cost per conversion
Spend divided by those conversions, returned by the same tool.
LSA spend
What Local Services Ads charged in the window, from get_lsa_summary.
LSA charged leads
The leads Google billed. Plain leads stay out, because spend buys only the charged ones.
LSA cost per charged lead
What each billed LSA lead cost on average. DuckDive gives the two inputs; your agent does the division.
GA4 conversions
Total key events from get_ga4_overview, all channels. DuckDive's tools say conversions; GA4 calls the actions that matter key events.
GA4 paid search conversions
The part GA4 credits to the cpc medium, from get_ga4_channels. It exists only to be subtracted.
Unpaid GA4 conversions
GA4 conversions minus paid search and any other paid medium: organic, direct, referral and email.
Blended cost per lead
Ads spend plus LSA spend, divided by Ads conversions plus LSA charged leads plus unpaid GA4 conversions.
Ask your agent
Build a ledger for every site: Google Ads spend and conversions, LSA spend and charged leads, GA4 conversions, all for the last 30 days.

Tools it calls list_websites get_google_ads_summary get_lsa_summary get_ga4_overview

By hand, that is three logins per site and a spreadsheet stale by Friday. DuckDive stores each site's Google Ads, Local Services Ads and GA4 history once linked, and whichever agent you already work in, whether Claude, ChatGPT, Cursor or a different MCP app, reads it a row at a time. It reads; it does not spend. Budgets still change in Google Ads and the LSA account, by your hand.

Before the first pull: three row checks

  1. Every row is a website. DuckDive files every number under a website, and each website takes a single Ads account and a single LSA account. So the sites you budget for are the websites you add, no more and no fewer. A location without its own website shares that website's row.
  2. Blank is not zero. A row whose LSA account was never linked comes back with nothing in both LSA cells. Write "not linked" there, so the ledger cannot be read as a site that spent nothing and won nothing.
  3. A half-imported row misleads. Before any division, have get_sync_status checked for every row. A source still importing can return partial history, which looks exactly like a site that cut its spend. Build the ledger once every source on every row reads ready.

Signing up with Continue with Google is enough to start the first row. To have your agent add the other sites and link their accounts, give it this line:

Give your agent this line
Set up DuckDive for my website. Follow https://analytics.duckdive.ai/agents.md

Which lead count belongs in which column

Most ledger fights are about one lead sitting in two columns, or a contact that never should have counted. Sort each kind of lead first.

Where each kind of lead lands in the ledger
LeadLedger columnKeep it out of
Form fill after a Google Ads clickAds conversions, if that form is a primary actionUnpaid GA4 conversions
Call from an ad, tracked as a conversionAds conversionsEvery GA4 column; no site visit happened
Call or message through an LSA adLSA charged leads, once Google bills itAds conversions, until that account moves into Google Ads
Form fill from organic search or a direct visitUnpaid GA4 conversionsBoth paid columns
GA4 key event credited to paid searchSubtracted from GA4 conversionsUnpaid GA4 conversions
GA4 key event imported into AdsAds conversions, only if no Ads tag counts the same formA second primary action
LSA lead Google did not chargeNowhere, spend did not buy itLSA charged leads

Sources: Google's View reports on Local Services Ads and Understand your conversion tracking data, checked November 2026.

What never reaches a column

Finance will ask for numbers the ledger cannot hold. Say so up front, and say where each one lives instead.

Figures outside the ledger, and where to find them
Figure someone will ask forWhy the ledger lacks itLook here instead
Cost per lead by campaignget_google_ads_summary returns one total per account (what those totals cover)Campaign reports inside Google Ads
Booked jobs, revenue, return on spendNo CRM, job or revenue source is connected, so the ledger stops at leadsYour CRM or job system, joined by date and site
LSA calls versus messages, LSA budget capsget_lsa_summary stops at three figures: spend, all leads, charged leadsThe Local Services Ads account
Meta or Microsoft Advertising spendDuckDive reads Google sources plus Yext and Local Falcon, nothing elseEach platform's own reporting
Totals tied to an invoice monthEvery window is rolling, never a calendar monthGoogle's billing pages
What next month's leads will beThe agent divides past figures and does not model auctionsA small budget test, read in the next ledger

So skip DuckDive if you run one site on one channel, need every number tied to an invoice month, or want budgets moved automatically. It pays off when several sites run both Google ad channels and the budget question returns every month. Teams with dozens of sites will find the wider routine on the page for operators running many websites.

Avoiding double counting between Ads and GA4

A visitor clicks an ad and requests a quote. The Ads conversion fires, and so does the GA4 key event, filed under paid search. Add the two together and one request counts twice. Four rules prevent it.

Rule 1: Ads owns the paid click

Keep the lead in Ads conversions and subtract GA4's paid search figure from GA4's total. Google's help on creating conversions from Google Analytics events draws the line: Ads conversion measurement counts only Google Ads sources, while a GA4 property can also count conversions from other sources.

Rule 2: one primary action per real lead

Inside Ads, a form counts twice if the Ads tag and an imported GA4 key event are both primary. Google's Analytics help on creating Ads conversions from GA4 key events says these are set to secondary "to prevent counting the same event twice for bidding." Flipping one to primary undoes that.

Rule 3: match the counting settings

Ads can count every conversion after a click or one per click; its page About conversion counting options recommends One for leads. GA4's key event counting method offers once per event or once per session, and a change applies to future key events only. Mismatched settings inflate one column against the other.

Rule 4: expect the paid columns to differ a little

Ads files a conversion under the click date; GA4 records it when it happens, and some events are filtered out of Ads as invalid clicks. So the two paid figures rarely match. Watch for a ratio that holds across sites; one site far off the pack has a settings problem.

Ask your agent
Line up each site's Google Ads conversions beside GA4's paid search conversions over the past 30 days. Which sites have a ratio unlike the others?

Tools it calls list_websites get_google_ads_summary get_ga4_channels

If a site's cpc row looks thin next to its Ads column, the GA4 integration page explains which visits GA4 files under each medium.

Worked example: a ledger for eight sites

Brackenridge Home Comfort is a made-up heating and cooling company with eight service-area sites. All run Google Ads; seven run Local Services Ads.

EXAMPLE DATAFictional business: Brackenridge Home Comfort

Build Brackenridge's ledger for all eight sites over the last 30 days, and split out GA4 paid search so nothing counts twice.

  1. list_websitesfinds the eight Brackenridge sites and their linked sources
  2. get_google_ads_summaryspend, conversions and cost per conversion, once per site
  3. get_lsa_summaryspend and charged leads for the seven sites with LSA
  4. get_ga4_overviewtotal conversions, once per site
  5. get_ga4_channelsconversions by medium, to remove paid search

Across the eight sites, $56,230 of paid spend sat against 1,371 leads, or $41.01 per lead blended. By channel, Ads averaged $71.09 per conversion and LSA $61.18 per charged lead. Thornhaven is the dearest site blended and Larkspur the cheapest. In the table, Ads is cost per conversion and LSA is cost per charged lead.

SiteAdsLSABlended
Westfield$80.00$55.00$45.10
Larkspur$60.00$90.00$36.18
Quarry Bend$75.00$50.00$40.38
Kestlemoor$90.00$40.00$37.50
Pewter Ridge$65.00$60.00$39.46
Fallowmere$70.00none$36.84
Thornhaven$80.00$65.00$50.80
Hatherleigh$50.00$90.00$39.73

Check Pewter Ridge before moving money on it. It logged 80 Ads conversions, but GA4 credited only 30 key events to paid search. Elsewhere, Ads runs 0.94 to 1.18 times the GA4 figure. Two primary actions are probably counting one form, so its $65.00 may be too kind. Fallowmere has no LSA account, so its cell is empty, not zero, and stays out of the LSA average.

Google Ads Local Services Ads Google Analytics 4

Next: where do the two Google channels disagree most?

Ask your agent
Using that ledger, which two sites have the widest gap between Ads cost per conversion and LSA cost per charged lead?

Tools it calls list_websites get_google_ads_summary get_lsa_summary

For Brackenridge it is Kestlemoor and Hatherleigh, with gaps pointing opposite ways. A gap is a reason to look, not a verdict: the cheaper channel may already be spending all it can.

Your own sites, one row each

Link each site's Ads, LSA and GA4 accounts once, and your agent can rebuild this ledger whenever the budget comes up.

Continue with Google

Free for 2 sites · No credit card

Blended versus channel cost per lead, location by location

Blended cost per lead answers "what does a lead cost this site, all in?" It belongs in the owner's report, and it flatters: strong organic traffic keeps it low even when paid dollars are wasted.

Channel cost per lead answers "what does a lead from this channel here cost?" That is the budget question. Ads cost per conversion and LSA cost per charged lead compare fairly because both divide spend by leads that spend bought.

  • Use blended for owner reports and to catch fading organic.
  • Use channel figures to move money between Ads and LSA at one site, or between sites in one channel.
  • Use neither on a channel with fewer than ten leads in the window, where one lead shifts the figure by around a tenth.

Check units before trusting a gap. A charged LSA lead is a contact Google billed; an Ads conversion is whatever the account's primary actions count. If one of those is a brochure download, cost per conversion is not cost per lead.

Ask your agent
Rank every site by blended cost per lead over the last 90 days, then show me which ones rank a lot better or worse on Ads or LSA cost per lead than they do blended.

Tools it calls list_websites get_google_ads_summary get_lsa_summary get_ga4_overview get_ga4_channels

Worked example: moving budget between two sites

The ledger shows averages. A budget move needs to know what the latest dollars bought, which a 90-day pull set against a 30-day pull shows.

EXAMPLE DATAFictional business: Brackenridge Home Comfort, Westfield and Larkspur

If I move 10% of Brackenridge Home Comfort's Westfield site budget to its Larkspur site, what does each site's last 90 days suggest would happen?

  1. list_websitesfinds the Westfield and Larkspur sites
  2. get_google_ads_summarythe 90-day window, then the 30-day one, for both sites
  3. get_lsa_summarythe 90-day window, then the 30-day one, for both sites

I can't see budget settings, so Westfield's last 30 days of paid spend, $12,900, stands in for its budget; ten percent is about $1,290. Each line compares the last 30 days with the monthly average of the 60 before.

Site, channelSpendLeadsPer added lead
Westfield Ads+$1,2000none
Westfield LSA+$300+3$100.00
Larkspur Ads+$600+9$66.67
Larkspur LSA+$300+2$150.00

Westfield's extra $1,200 in Google Ads bought nothing: 120 conversions at $8,400, and 120 again at $9,600. That slice is the cheapest money to take back. Larkspur's added Ads conversions came in close to its $60.00 average, while its added LSA leads cost far more than its $90.00 average.

So the last 90 days point to one move: Westfield's Google Ads to Larkspur's Google Ads, with both LSA accounts left alone. That reads past ratios; it is not a forecast. Auction prices, weather and Larkspur's budget cap can change the outcome, and I can't see them. Confirm Larkspur's campaigns aren't limited by budget, then rerun this in 30 days.

Google Ads Local Services Ads

The agent did not promise a lead count. Per-added-lead figures rest on small changes: a direction to test, graded by the next ledger.

Rules for when to move money

Before money moves between sites or channels

  • Both windows agree. The 30-day and 90-day ledgers point the same way. If they split, wait a month.
  • Enough leads to judge. Ten or more in the window for each channel on both sides of the move.
  • No open flags. Every row involved passed the double-count check.
  • Marginal beats average. Take money where added spend bought least, not where the average is highest.
  • The receiver can spend it. The receiving campaign or LSA account is not already capped.
  • Same unit on both sides. Primary Ads actions are real leads, not page views or downloads.
  • One move at a time. Shift one slice, change nothing else at either site, and read it after a full window.
  • Late conversions first. Ads files conversions by click date, so judge a move only once its window has closed.

The cap check matters most for LSA. Google's About Target CPL bidding page says Local Services Ads "will never spend beyond your set monthly budget," and DuckDive does not read budget settings. Prices also move with rivals: How bidding works for Local Services Ads says competing local bids set what a lead is worth.

Ask your agent
Run my reallocation checks on every site for the last 30 and 90 days: flag channels with under ten conversions or charged leads, Ads conversion counts out of line with GA4 paid search, and sites where the two windows disagree on which channel is cheaper.

Tools it calls list_websites get_google_ads_summary get_lsa_summary get_ga4_channels

Edge cases in the ledger

Decide these cases in advance and note each decision under the table.

  • No LSA account. Leave its cells blank, never zero. A zero drags the group's LSA average down and fakes free leads.
  • Spend with no leads. Cost per lead is undefined, not zero. Put those rows at the top.
  • Other paid platforms. If Meta or Microsoft clicks are tagged with a paid medium, GA4 files them outside cpc and the unpaid column already leaves them out. Note their leads under the table, because the spend that bought them never reaches the ledger.
  • One Ads account for two sites. DuckDive links one Ads account to a website and reports it as one total, so shared spend cannot be split. Ask the agent for each site's paid search (cpc) sessions from get_ga4_channels, link the account to the site with more, and footnote both rows.
  • Counting changes mid-window. GA4 does not rewrite past data after a counting switch. Note the date and compare only windows wholly on one side of it.
  • A month that keeps growing. Ads files late conversions back under the day of the click, so the month just ended gains leads for a while. Before it is judged, an admin can ask the agent to pull up to 30 days of a site again; DuckDive allows that once per website every 15 minutes.
Ask your agent
Which sites spent money on Google Ads or Local Services Ads in the last 30 days but recorded no conversions or no charged leads?

Tools it calls list_websites get_google_ads_summary get_lsa_summary

An LSA account that moved into Google Ads

Google's transition help page says that after the move, the Conversions metric "displays the total number of leads for which you're charged," and Cost / conv. is your average cost per lead. So a migrated site's charged leads can turn up in the Ads columns. In the month of the move, check both columns for the same leads before adding. The guide to saving LSA history before the move covers what to keep.

Many locations on one domain

DuckDive organises data by website, so a single domain carrying a page for each location is still one row; the multi-location overview shows how to split it. If leads are vanishing rather than getting dearer, start with finding which locations lost calls. When a storm sends costs swinging weekly, the roofing storm-season playbook shows how to read short windows.

Questions from whoever signs off the budget

Can I build the ledger for a calendar month instead of the last 30 days?

Not through DuckDive today. Each pull is a rolling window, anywhere between a week and a year, and no tool takes a start date and an end date. Run the 30-day ledger early in a month for a close monthly view, and use Google's billing pages when a figure must match an invoice exactly.

Does DuckDive hand me cost per lead, or does my agent work it out?

It depends on the column. get_google_ads_summary returns Google's own cost per conversion for the account. Local Services Ads returns spend and charged leads, so your agent divides them. Blended cost per lead is always the agent's arithmetic, so ask it to list spend and lead counts beside each ratio.

Can my agent move budget between sites once it finds the cheaper one?

No. DuckDive only reads from the ad accounts; no tool it offers can touch a bid, a budget or a campaign. The agent can tell you where recent spend bought the fewest extra leads. You make the change in Google Ads or the LSA account.

Does the ledger need Search Console linked?

No. The ledger reads Google Ads, Local Services Ads and GA4. Link Search Console anyway if you want your agent to explain why a site's unpaid conversions moved, because it shows the same site's organic clicks, queries and pages.

Will the ledger match what Google bills me?

Spend comes from the same accounts, so it should land close, but expect small gaps. Windows roll instead of following billing months, Ads files conversions under the click date, and LSA lead credits Google grants can change what a period cost. Before the figures go to finance, have an admin queue a fresh import of the past 30 days, then rebuild the ledger.

Can I try the ledger on one site for free?

Yes, and the first site asks for no payment details. That single site already fills a full ledger row, with Google Ads, Local Services Ads and GA4 side by side and the double-count check applied. Add sites as the ledger grows.

Sources

  1. About key events, Analytics Help, checked for this guide on November 16, 2026.
  2. View reports on Local Services Ads, Google Local Services Help, checked for this guide on November 16, 2026.
  3. Understand your conversion tracking data, Google Ads Help, checked for this guide on November 16, 2026.
  4. Create conversions from Google Analytics events in Google Ads, Google Ads Help, checked for this guide on November 16, 2026.
  5. Create Google Ads conversions based on Google Analytics key events, Analytics Help, checked for this guide on November 16, 2026.
  6. About conversion counting options, Google Ads Help, checked for this guide on November 16, 2026.
  7. Change the counting method of key events, Analytics Help, checked for this guide on November 16, 2026.
  8. About Target CPL bidding, Google Local Services Help, checked for this guide on November 16, 2026.
  9. How bidding works for Local Services Ads, Google Local Services Help, checked for this guide on November 16, 2026.
  10. Local Services Ads transition to Performance Max campaigns with pay-per-lead goals, Google Ads Help, checked for this guide on November 16, 2026.

Last reviewed . Numbers in EXAMPLE DATA blocks are fictional.

Count every lead once, on every site

Link each site's Google Ads, Local Services Ads and GA4, then ask your agent for the ledger whenever the budget is on the table.

Free for 2 sites · No credit card · Working in an agent? Let it set DuckDive up