Dissecting the data reveals big tech’s wasteful secret

AI advertising systems are masking some startling inefficiencies, says AccuraCast’s Farhad Divecha. It’s time to weed them out with a closer look at the data.

The most powerful tech platforms are good at getting solid results for advertisers, but they don't always identify how brands could do even better.

Step in AccuraCast, the global digital marketing agency seeking to plug the sector’s knowledge gaps by getting under the hood and delivering insights rarely shared by the likes of Facebook, Google and Amazon.

In an interview with Performance Marketing World, the business’s managing director, Farhad Divecha, explains how brands should strategically prune certain ads or online products to optimise investments when using AI-driven platforms.

“Automation will always compensate for underperforming sales,” Divecha says. However, he warns the discrepancy has a tendency to be hidden from advertisers in the marketing portfolio.

“For example, we have a client who spends a lot of money advertising across a lot of different markets, and when we look at Facebook or Google, we see they’re working brilliantly. But when we dissected the numbers and looked into the minutiae, there was actually 50% of adspend going to absolute duds.”

The problem, Divecha says, is that there is “very little incentive” for a platform or an AI which works on a portfolio basis to balance adspend accordingly.

“They have their money,” he says, and so their concern ends there. Yet it is limiting growth and denying brands the promise of true performance in their online marketing. 

“And it doesn't just apply to product ranges,” Divecha says. “This problem goes on to audiences, creative, targeting; everything. If you don't dive into the data, there's a very high likelihood automation will compensate.”

Although marketing bosses may feel content with their performance advertising results, those seeking to break into new growth cycles are likely to find efficiencies simply by more thoroughly interrogating the data.

“If really good [ads] compensate for the bad ones, you still get net good results,” Divecha says. 

“You’re happy, your CMO is happy; everybody is happy because you're profitable. But what no one realises is, that if you want to grow, you have to start killing off all these bad [ads].”

The haves, and have-nots

Divecha, who has been with the business for 18 years, and works with brands including Penguin, Waitrose and Center Parcs, also warned the sector risks creating a two-tier system as data privacy becomes a driving force online. 

Referring to Apple’s iOS 14.5 update, which allows users to stop advertisers tracking them through apps, Divecha says small businesses were being disproportionately penalised.

“Ten years ago, digital used to be a level playing field,” he says. “But what Apple has done is basically wreck the place and made it worse for small businesses.”

With Google also set to deprecate third-party tracking cookies by the end of 2023, the commercial internet is undergoing a vast transformation as platforms and advertisers seek alternative methods of tracking.

However, Divecha says those with the most money will gain the upper hand by essentially being able to pay to feed first-party data in through the “back door” at the major platforms, while smaller advertisers, with less resources and technical expertise, will need to rely more on attribution modelling.

“It's a big problem,” says Divecha. “And in the long run, it'll become a case of the rich getting richer, because they can invest in the single customer views and the big, single data management platforms.

“For the small businesses, they're going to lose visibility as they rely more on attribution modelling, based on what they think is actually happening...and that doesn’t sit well with me.”

The loss of tracking cookies also means brands are starting to lose sight of their sales. With platforms such as Mozilla or Apple’s Safari already blocking third-party tracking cookies as default, and Google soon to follow, Divecha says some advertisers were only seeing 80% actual sales conversions. 

This was resulting in the unnecessary “throttling” of campaigns that might appear to be underperforming, when in reality the brand was just blind to the sales coming through.

“The AI systems think you're performing less than you really are. If you have a fixed budget, that means you're telling the system it needs to overcompensate when it actually doesn't.”

For brands, that might mean holding their nerve and operating with some blind spots as the industry sets about recalibrating the way it works with less reliance on tracking data. 

Yet longer-term, it will mean doing more with less data - so making the right investments to understand the true picture will be crucial, suggests Divecha. 

“AI is only as good as the data,” he wants to remind advertisers, particularly those looking to scale up internationally and who are assessing ways in which to tighten up their performance channels.

“And for us, it’s always been about that; about using data to see where the opportunities are to make things happen, and make them happen quickly.”

Farhad Divecha is the managing director at AccuraCast