Paid Search Guide: Google Ads, Meta Ads and Measurement

Start with the decision in front of you: the budget, the Google Ads setup, or the measurement fix. Read the pages for that goal in order, because each one builds on the decision made in the page before it.

If you already run campaigns, start with Goal 3 and check the conversion data before you change any bid or budget. A broken conversion feed distorts every later decision.

Size your first test from your own cost per lead, not from an industry average. Use the averages below to check whether your numbers fall in a plausible range.

LocaliQ, with WordStream, reports data from thousands of its customers' Google Ads and Microsoft Ads campaigns. It reports an average cost per click of $5.42 for search advertising across all industries.

The same report gives an average click-through rate of 6.64% for search advertising across industries.

For Google Ads alone, the report gives an average conversion rate of 8.18% across industries. Do not divide the cost per click by that rate to estimate a cost per lead, because the two averages cover different scopes.

The report also gives an average cost per lead of $66.69 for search advertising.

Set your test budget from your target lead count times your own cost per lead. Compare that figure with these averages only when your metric definitions match theirs.

Read the click-through rate and the cost per click together. A weak click-through rate points at the ad copy. A high cost per click, with a click-through rate in line with your own history, points at bidding and keyword competition.

Work backward from the value of a lead. The cost per lead you can afford, together with your own conversion rate, sets the most you can pay per click. Compare that ceiling with the averages above before you commit to a budget.

Benchmarks from other vendors can differ widely, so treat any single average as a starting point. The paid search strategy page covers search economics and how to split a B2B budget across campaigns, so read it before you set the test budget.

Confirm that conversions reach the ad platform inside the import window before you raise spend. A late or missing conversion can make a working campaign look weak.

Before you raise spend, send one test conversion for a known CRM record. Check that it lands in the account with the conversion time you expect, then scale the upload.

Map each conversion action to one CRM stage, such as a qualified lead or a closed sale, and write the mapping down. Changing the stage in the middle of a test breaks the comparison between campaigns.

Google Ads accepts an offline conversion uploaded with a GCLID up to 90 days after the click.

Conversions uploaded with personal data for enhanced conversions have a shorter limit of 63 days. Google does not import uploads made after the limit.

Google Ads does not import the same conversion more than once. Conversions that share a click count separately only when their times differ and conversion counting is set to every.

Imported conversion statistics take time to show up in the account. Validate them with the all conversions by conversion time column before you judge a campaign on them.

Data-driven is the default attribution model for most conversion actions. First click, linear, time decay and position-based models are no longer supported, and last click still is. Choose between data-driven and last click, and keep the same model in every budget report.

For users in the European Economic Area (EEA), Google requires consent collection and consent signals shared with Google to keep using tags for measurement, ad personalization and remarketing.

If tags load only after the consent banner is answered, Google cannot verify the consent choices, and data may be lost. Test how your banner and tags load on a staging page before you launch.

Google Ads says that starting in April 2026 it accepts user-provided data from website tags, Data Manager and API connections together, under a single on or off setting for enhanced conversions for web and leads. Check which setting your account uses before you change tags, so the data keeps arriving.

The offline conversion tracking page covers the GCLID upload, enhanced conversions and the attribution choice in depth.

Decide whether Performance Max fits the account before you build asset groups. Confirm the conversion import first, because the campaign can only work from the conversions it receives.

Choose the conversion action before you choose the bidding approach. The bidding works from that action, so a vague action produces vague results.

Performance Max ads are served on the following channels: Search, Maps, YouTube, Discover, Gmail and the Google Display Network.

Channel-level data is in the Performance Max channel performance report. Read that report before you cut a campaign, because spend and conversions can sit on different channels.

For account setup, negative keywords and bidding, use the PPC campaign management page.

The Google Ads guide covers asset groups, competitor traffic capture, omnichannel retargeting and testing. Read it when you plan the whole campaign set, not just one campaign.

Decide whether a Meta change deserves a test before you launch it, and set the success measure first.

Meta describes Andromeda as its proprietary machine learning system for retrieval in ad recommendation.

In that engineering post, Meta reported a +6% recall improvement to the retrieval system and a +8% ads quality improvement on selected segments.

Those gains describe Meta's system, not your account. Read your own cost per result before and after each change, and keep the change only when that figure improves.

Compare ad sets only on the same conversion event and the same spend window. Do not judge a new ad set on its first clicks, because the early numbers move before the ad set settles.

The Meta Ads guide covers the profitability equation, available customer acquisition cost and the ad selection pipeline. Start there when you set the Meta budget.

Test the ad copy and the landing page in separate experiments.

Change one variable per test, such as the headline, the offer or the first screen of the landing page, and keep the rest fixed. Write the hypothesis from the ad copy frameworks before you launch.

Use the landing page to answer one question at a time: whether the offer converts, or whether the form loses people. A single page change cannot answer both.

The ad creative and CRO page covers ad copy frameworks, a swipe file, creative testing, video creative and landing page optimization.

When you judge a new landing page, use cost per lead as the measure, not click volume. A page can gain clicks and still lose leads.

Open the page that answers the decision in front of you, then use the other pages for the next decision.

The SEO fundamentals page includes a panel on when to prioritize organic search over paid, so read it if the budget question is still open.