In plenty of paid search accounts, a loyal customer is typing the brand’s name into Google, clicking the ad sitting above the organic listing and buying what they came for, while the dashboard logs it as an acquisition. One CAC number can describe a brand paying more to win people who have never bought from it, or a brand paying Google to resell itself to customers it already had.
Ashwin Palo is Assistant Director of Product and Growth Marketing, Data Analytics at HSBC, where his work has included building an in-house multi-touch attribution platform. Before HSBC, he worked on the agency side at Webenza, as a Senior Account Coordinator at Google, and wrote Uncluttering Programmatic Advertising.
“When CAC is going up, you need to check whether you’re acquiring the same customer or getting a new customer. If it’s a new customer, it’s fine. If it’s the same customer, it’s not good,” Palo says. Plenty of performance reports would sail through a quarterly review and still misdescribe the business. They count the month’s sales and leads with no way of knowing how many of those buyers were already customers, so the standard CAC formula books a loyal shopper’s fourth order and a stranger’s first as the same win. Separating new buyers from returning ones requires customer-level data. Establishing whether the advertising brought those buyers in at all requires a different test.
Brand search bills for customers already won
Branded search is where the confusion gets most expensive, because the people searching for a brand by name are often its own customers. “If you’re running ads on your brand campaigns, your ad will be there on the top. They’ll click it, and since they’re already your customer, they’ll do the transaction with you. To get the same sale from the same customer, you’re spending more money,” Palo explains.
The campaign reports a strong return and the platform books a conversion, while the business pays again for a customer relationship it already built. Palo makes one exception, for customers who have asked a brand to stop sending marketing emails and texts. Paid media can give the brand another way to reach those buyers, which makes that spend easier to justify.
Palo’s fix starts with an ordinary month of 100 sales. Each sale gets tagged with the channel that received credit for it, from organic search and owned channels like WhatsApp through to paid, even though a single journey often passes through several. That tally shows how much of the month paid media claims, and it frames whether CAC should be read on paid spend alone or across the blended business. The paid sales then get split by customer status, product by product, and the number worth watching is the share coming from people who had already bought. For campaigns meant to win new customers, “the lower the percentage you keep for existing customers, the better it is,” he says. Repeat purchases remain welcome in most businesses. The trouble starts when they get counted as proof that paid media is winning new customers.
Each platform’s report argues for more of itself
A platform can report every conversion it saw accurately and still overstate the business it created. Every platform evaluates performance through its own measurement framework, with its own attribution windows and conversion definitions, so its reported contribution can’t serve as a neutral account of the mix. “If you’re a Meta-heavy brand, then regardless of whatever you do, Meta will say its awareness is giving you more new customers, so you should run more ads on its platform. If you’re Google-heavy, Google will also bias you,” Palo says.
Meta added an incremental attribution setting to its options, and it’s the only one built on modeled data, which leaves advertisers taking the platform’s word for what counted. A platform’s “new customer” can simply be a buyer that platform hadn’t converted before, even when the brand has been selling to them for years. An independent view helps by reconciling channel reports against the brand’s own customer and sales data. When a shopper buys a T-shirt from an online retailer, the retailer can trace the products browsed beforehand and match that behavior against the paid-media interactions it can observe, month by month. “On Meta I’m running 10 campaigns, and these 10 campaigns have driven 1,000 people to the website, whether they bought or not,” he says.
That view shows which journeys paid media touched. Proving the touch caused the purchase takes an experiment, such as a holdout that pauses ads for part of the audience, and three in four marketers say their current measurement isn’t delivering the accuracy and trust they need. For smaller teams, Palo points to Microsoft Clarity, a free tool that shows how a campaign’s visitors behave once they land. It helps diagnose a campaign, and the question of whether the campaign worked stays with testing.
Reach pays off when it adds customers
The same journey data gives Palo a first read on awareness spend. He looks at how much converting traffic passed through top-of-funnel campaigns and weights the budget from there, with no category ratio involved. “Every business is unique, and every business is at a specific level,” he says. That share works as a hypothesis about where awareness is pulling its weight, and the budget should follow whatever testing confirms.
A campaign can reach millions of people while mostly finding the ones who already know the brand. One warning sign is reach climbing while new-customer counts stay flat and frequency rises, suggesting that more of the budget may be going toward people the brand has already reached. Expansion earns its budget through incremental revenue and new customers who wouldn’t have arrived otherwise, which is why Palo prefers to grow carefully. “Always, when you expand, expand slow,” he says.
Every new platform arrives with its own dashboard and its own case for more budget. A quarter spent optimizing toward sales that would have happened anyway can look excellent on every one of those dashboards, and the customer count at the end of it barely moves.
