PPC Attribution: Understanding What Actually Converts

Attribution, the question of what actually caused someone to become a customer, sounds like it should be simple. Someone clicked an ad, they became a customer, the ad gets the credit. Job done.

In reality, it is messier. Someone might see your Facebook ad, ignore it, later search for your service on Google, click an ad, browse your website, leave without enquiring, come back directly a week later, and then call you. Which touchpoint deserves the credit? The Facebook ad that planted the seed? The Google ad they clicked? The direct visit when they finally converted?

This is the attribution problem, and every service business running PPC needs to understand it, even if only to avoid making bad decisions based on misleading data.

Why Attribution Matters

Attribution matters because it affects where you spend your money. If you think Google Ads is generating all your leads and Facebook is generating none, you will shift budget to Google. But what if Facebook is creating awareness that Google is then capturing? Cut Facebook and you might find your Google performance drops too.

For most local service businesses, the stakes are not enormous. You are not running million-pound campaigns where a few percentage points of attribution error cost tens of thousands. But understanding the basics helps you avoid obvious mistakes and set realistic expectations about what your data is telling you.

How Attribution Models Work

Attribution models are rules for assigning credit to different touchpoints in a customer journey. Different models give different answers to the same data.

Last-click attribution

Last-click attribution gives 100% of the credit to the final touchpoint before conversion. If someone clicked a Google ad, then came back directly, then called you, the direct visit gets all the credit. The Google ad gets nothing.

This is the simplest model and was the default in most advertising platforms for years. It is easy to understand but systematically undervalues awareness-building activities that happen earlier in the journey.

First-click attribution

First-click attribution does the opposite, giving all credit to the first touchpoint. The Facebook ad someone saw three weeks before they converted gets the credit, even if they clicked five other things since.

First-click overvalues top-of-funnel activities and ignores everything that happened between first touch and conversion. It is rarely useful on its own.

Linear attribution

Linear attribution spreads credit equally across all touchpoints. Four touchpoints each get 25% credit. This feels fair but treats a casual ad view the same as the final click that drove the conversion.

Time-decay attribution

Time-decay gives more credit to touchpoints closer to the conversion. The touchpoint immediately before conversion might get 40% of the credit, while something three weeks earlier gets 5%.

This is often a reasonable model for service businesses. Recent touchpoints probably did more to drive the actual conversion than distant ones.

Position-based attribution

Position-based, sometimes called U-shaped, gives most credit to the first and last touchpoints (often 40% each), with the remainder split among middle touchpoints.

This acknowledges that first impressions and final conversions matter most, while still giving some credit to the journey in between.

Data-driven attribution

Google Ads and some other platforms now offer data-driven attribution, which uses machine learning to analyse your actual data and determine how much credit each touchpoint deserves. This sounds sophisticated and often is, but it requires significant conversion volume to work properly. Most local service businesses will not have enough data for data-driven models to be reliable.

Cross-Platform Attribution Challenges

The real attribution headache comes when you advertise across multiple platforms. Google, Meta, LinkedIn, and Microsoft all want credit for your conversions. And they all measure in ways that favour themselves.

Each platform tracks its own touchpoints but has limited visibility into what happened elsewhere. Google can see the Google ad click, but it cannot see that the person saw three Facebook ads first. Meta sees the Facebook impressions but does not know about the Google click that followed.

This leads to over-counting. If someone saw a Facebook ad and later clicked a Google ad before converting, both platforms will claim the conversion. Add up all the conversions each platform reports and you will get a number larger than your actual total.

There is no perfect solution to this. Accept that platform-reported conversions will not match your actual lead count. Use each platform's data to optimise within that platform, but look at your actual business results to judge overall effectiveness.

Privacy Changes and Their Impact

Attribution has become harder over the past few years due to privacy changes. Apple's App Tracking Transparency, introduced with iOS 14, lets users opt out of cross-app tracking. Most users do opt out.

For advertisers, this means Meta, in particular, has much less visibility into what happens after someone clicks an ad. Conversion data is delayed, modelled, and often incomplete. Google has similar challenges with increasing browser privacy restrictions.

The practical impact: do not trust platform conversion numbers as absolute truth. They are directionally useful but imperfect. A platform reporting 20 conversions might mean you got somewhere between 15 and 25 actual leads. The trend matters more than the exact number.

Setting Up Proper Conversion Tracking

Imperfect attribution is no excuse for not tracking at all. Basic conversion tracking is essential. Without it, you genuinely know nothing.

Google Ads conversion tracking

For Google Ads, set up conversions for every meaningful action: form submissions, phone calls from your website, clicks on email addresses, and calls from call extensions.

Form submissions are usually tracked via a thank you page. Someone submits the form, lands on a page that only appears after submission, and that page load triggers the conversion.

Phone calls can be tracked using Google forwarding numbers. For calls from ads, this works automatically with call extensions. For website calls, you need to implement website call conversion tracking, which swaps your phone number for a Google tracking number for visitors from Google Ads.

Meta conversion tracking

For Meta, install the Meta Pixel on your website and configure conversion events. Track the same actions: form submissions, phone calls, key page views.

Meta's Conversions API can improve tracking accuracy by sending conversion data server-side rather than relying solely on browser-based tracking. This is more complex to set up but increasingly important as browser tracking becomes less reliable.

Google Analytics

Google Analytics sits above individual ad platforms and can give you a more complete picture. It tracks all traffic sources to your website and can show you the full journey a visitor took across multiple sessions.

Set up goals in Analytics that match your conversion events. Then use the multi-channel funnel reports to see how different channels contribute to conversions. This will not be perfect either, but it provides a useful cross-platform view that individual ad platforms cannot.

Understanding Customer Journeys

Before optimising based on attribution data, understand what customer journeys actually look like in your business.

For emergency services, journeys are often short. Someone has a blocked drain at 8pm. They search, they click, they call. Single touchpoint, immediate conversion. Attribution is straightforward.

For considered purchases, journeys are longer. Someone thinking about a new kitchen might spend months browsing, comparing, saving ideas. They might see your ad on Instagram, visit your website, leave, come back via Google, look at your gallery, leave again, and finally submit an enquiry weeks later. Multiple touchpoints, extended timeframe. Attribution is murky.

Most service businesses fall somewhere between these extremes. Understanding your typical journey helps you interpret attribution data sensibly. If you know that most customers interact with you multiple times before enquiring, do not expect any single channel to claim full credit.

Making Budget Decisions With Imperfect Data

Here is the uncomfortable truth: you will never have perfect attribution data. Every model has flaws. Every platform overcounts. Privacy changes have made tracking harder. You have to make decisions anyway.

Focus on directional trends

If Google Ads cost per lead is trending upward while lead volume is flat, something is changing. Maybe competition has increased, maybe your Quality Scores have dropped, maybe there is a seasonal factor. The exact conversion count matters less than the trend.

If Meta is showing good cost per lead but you are not seeing the leads come through to your CRM, something is off. Either tracking is broken or the reported leads are not real enquiries. Investigate before spending more.

Compare channels at the source of truth

Your actual business results are the ultimate source of truth. How many leads did you get this month? How many became customers? What was your revenue? These numbers do not lie.

If you are spending £1,500 across Google and Meta and generating 30 actual leads per month, your true cost per lead is £50, regardless of what the platforms report. If Google reports 20 leads and Meta reports 25, they are over-counting by about 50%, which is typical.

Use this reality check to calibrate your expectations. Platform data helps you optimise within each platform. Business data tells you whether the overall investment is working.

Run incrementality tests when possible

The gold standard for understanding attribution is incrementality testing. Turn off a channel for a period and see what happens to your overall results.

This is scary because you might lose leads. But if you suspect a channel is over-claiming credit, it is the only way to know for sure. Turn off Meta for two weeks. Did your Google leads stay the same, increase (because Meta was getting credit for journeys that would have happened anyway), or decrease (because Meta was genuinely driving awareness)?

Not every business can afford to run these tests, especially if lead volumes are already thin. But if you are spending significant money and unclear whether a channel is actually contributing, it may be worth the experiment.

Ask your customers

The simplest attribution method is often overlooked: ask people where they heard about you.

Add a field to your enquiry form: "How did you find us?" Include options for Google, Facebook, recommendation, and other. Train whoever answers the phone to ask the same question.

Self-reported attribution has its own flaws. People forget, misattribute, or say "Google" when they mean your website. But it provides a human check on your tracking data. If everyone says they found you on Google but Meta claims half your conversions, something is not adding up.

Tools for Attribution

Several tools can help make sense of attribution data.

Google Analytics 4 is free and provides multi-channel funnel reports showing how different sources contribute to conversions over time. The Model Comparison tool lets you see how different attribution models would credit the same conversions.

Google Ads Data Hub and Meta Advanced Analytics offer deeper attribution analysis but require significant technical setup and are overkill for most service businesses.

Call tracking platforms like Infinity or CallRail can track phone conversions across channels and provide source attribution for calls, which is valuable if phone enquiries are a significant portion of your leads.

CRM systems with proper source tracking let you follow leads through to customers and see which marketing sources generate revenue, not just enquiries. This is the closest you can get to true attribution.

A Practical Approach for Service Businesses

Here is a realistic attribution approach for a typical service business running Google and Meta Ads.

Set up conversion tracking properly on both platforms. Track form submissions, phone calls, and any other meaningful actions. This is non-negotiable.

Install Google Analytics and configure the same conversion events. Use it as a cross-platform view.

Ask every lead how they found you. Record the answers consistently.

Accept that platform numbers will over-count. When Google says 25 leads and Meta says 15, expect your actual total to be somewhere around 25 to 30, not 40.

Judge channels by business impact, not platform metrics. If spending £1,000 per month on Meta and your total leads stay the same, Meta is probably not contributing much, whatever their dashboard says.

Focus on trends rather than absolute numbers. If cost per lead is rising, investigate. If lead quality is dropping, investigate. The specific numbers matter less than the direction.

Do not over-optimise based on attribution data you do not trust. If you are unsure whether Meta is contributing, maintain modest spend rather than cutting entirely based on potentially misleading data.

Attribution is genuinely difficult, and the industry has not solved it. Accept imperfection, track what you can, verify against reality, and make decisions based on overall business results rather than platform-reported metrics alone. That pragmatic approach will serve you better than chasing perfect attribution that does not exist.