How the Facebook #stophateforprofit boycott presented the perfect opportunity to test the true impact of social media ads – with surprising results
SPONSORED BY Crimtan
How many times have you found yourself targeted on social media by ads for products you’ve just bought? Annoying, isn’t it?
It’s just one instance of how the algorithms used by social media platforms can get it very wrong. In many cases, these ads have been sold to companies as part of a costly digital paid social marketing campaign. Paid social was the main source of Facebook’s $70-billion revenue in 2019, which accounted for just over a fifth of estimated total digital ad spend.¹
Yet there is growing concern among savvy marketeers that the analytics used to plan and execute campaigns on social media are inherently weak. That matters because it means you could be spending money putting adverts in front of customers who were going to buy anyway. In other words, you might be wasting precious budget firing at the wrong target.
You could be spending money putting adverts in front of customers who were going to buy anyway. In other words, you might be wasting precious budget firing at the wrong target
Determining whether or not that’s the case isn’t easy. But action taken this year by more than 1,000 advertisers on Facebook has provided a unique opportunity to examine what might happen if you cut back on paid social. The results of our study were so surprising we had to run the numbers several times to prove it to ourselves and might just make you think again about whether paid social is a good use of your marketing spend.
But first, some context. Any marketing plan has to start with understanding the customer life cycle, which is easier said than done. Marketeers have focused on clicks as a measure of engagement and more specifically on the final press of the mouse that signals a purchase, so-called last-click attribution.
It’s a perfectly good measure, but it focuses only on the last couple of minutes of the customer’s journey. Most of us, depending on the product, will have done some research before we buy; maybe asked friends, read reviews, used a price comparison site or seen an advert somewhere else. We could have seen multiple ads on different channels and in different formats before deciding. We might even have browsed in a bricks-and-mortar shop.
Last-click attribution fails to capture the whole of that journey, the hard work of brand building and demand generation that goes on away from paid social. Yet it is important for marketeers to be able to work out the effectiveness of marketing spend and, more particularly, incremental return on investment, which is the difference between doing or not doing any particular part of the journey. That’s a tricky thing to measure without a parallel universe.
In July, more than 1,000 companies boycotted Facebook as part of a campaign putting pressure on the social media platform to change its handling of hate speech and misinformation.² #stophateforprofit called on businesses to pause advertising for a month and major brands from adidas to Verizon did just that. Some are still continuing the boycott.
This gave a unique opportunity for Crimtan to conduct the perfect A/B test and assess the difference on underlying revenue of Facebook being on or off. Across our clients, there is significant amounts spent on paid social activity, the majority being split between Facebook and its sister company Instagram, with the aim of leveraging potential customers who show an interest in their products. Many companies are unwilling to test the efficacy of that spend and risk losing sales.
But the boycott was the perfect parallel universe, a chance to assess the impact of Facebook and whether the money that is pumped into advertising on its platform generates a true incremental return, or simply puts ads in front of customers that were going to buy anyway, generating no incrementality.
We compared sales for the month of the boycott with a prediction of what sales would have been had brands continued to spend with Facebook, using the last three years of Google Analytics data to produce a prediction.
Astonishingly, switching off Facebook had no discernible impact on bottom-line revenues for the brands in the dataset. Indeed, the modelling suggested a modest improvement in underlying revenue as a result of turning off Facebook, but this was ruled out as being a significant upward trend.
Of course, one swallow does not a summer make. But the result feeds into our concern that basing marketing spend on current tracking methods, such as last-click attribution, is not giving marketeers a real picture of what is working well. It is looking at an unhealthily narrow metric, focusing on the easiest element to measure rather than what is most effective.
With the growing importance of customer privacy, social media platform analytics are increasingly obscure, with data carefully guarded. Yet for a quarter of the digital advertising spend market, which is money taken by Facebook and Amazon, you cannot track customer engagements using any tools other than those provided by these businesses.
Attribution should not aim to provide the perfect picture of your customer. Instead it should offer the best picture of the combinations that drive the best engagements and returns for your business. Customer actions, such as scrolling through pages or length of time spent on your website, are as important as last click, yet few marketeers have learnt to understand and challenge the finer details of social media analytics or statistical modelling. As a result, the billions spent on paid social could be money down the drain.
Marketeers need a means of analysing and measuring the whole of the customer’s digital journey. It is this deeper understanding that is provided by Crimtan’s hybrid attribution model and life-cycle product.
Analytics should be transparent in showing what works for you and your brand. Marketeers should be given the correct information to be able to deliver cohesive marketing strategies to find new customers, prospect them to conversion and grow lifetime value. It’s time to question whether the most popularly used metrics provide the answers you need.
To download the full white paper Is Paid Social Marketing Pointless? please visit crimtan.com