embrace SM

The Future of Social Media

The Future Of Social Media Is Embracing Experimentation

Total Media's head of digital performance discusses AI influence, the short-form video revolution, and adaptable measurement models

By Oli Williams

It’s fair to say that both advertisers and consumers have been spoilt for choice as to where to spend their budgets and time in recent years.

But whilst consumers face simpler choices (such as “trying” Threads before reverting to X - until Elon decides he wants to charge for whatever he’ll be calling the platform that month), advertisers face a feature buffet from every social platform. So how do advertisers ensure the right choice is made at the social feast?

The first thing to consider is undoubtedly AI - and not the “is it going to replace me and render me unemployed" kind of AI, but the Machine Learning kind. Whilst Google has long given advertisers the opportunity to take advantage of machine learning solutions in both the Search and Programmatic space for bidding or creative purposes, social platforms have taken a little longer to adopt this way of working. Somewhat predictably, many of the releases in this space come from Meta which now offers machine learning solutions for budgeting, creating and audience testing purposes through their “advantage” suite of products.

Many activation teams will have rejoiced at the thought of manual tasks being handled by a machine. However, making this shift comes with a trade-off in control over audience selection as platforms preach the "broad is best" narrative; advertisers have always been reluctant to align to this and potentially other recommendations that see them relinquish control. So in response, advertisers should undoubtedly continue testing before adopting certain approaches fully. Thankfully a lot of AI-led innovations replace existing ways of doing things and can be solved through A/B and Single Cell Testing.

But is an innovation that requires a more complex solution to testing really an innovation at all? For all the innovation in the social space, it is somewhat surprising to see that many of the new approaches are simply “duplication” rather than "innovation". None more so than the various social platforms' solutions to short-form video.

It goes without saying that the catalyst for this is TikTok’s success in recent years and what looks like “duplication” from the likes of Instagram with Reels and YouTube with Shorts are actually conscious efforts to take advantage of the ever-changing landscape. With 24 per cent of people now turning to short-form video as their primary search engine, social platforms want to give advertisers, in particular, the opportunity to reach the consumer at the point that is often considered closer to the point of conversion. With this comes increased advertising spend going to social platforms as they start to chip away at budgets that may have been traditionally allocated to paid search.

So how can advertisers ensure they are making the right decision when allocating media spend to these formats? The key is thinking about short-term video as the channel and optimising spend accordingly, the way you would think about allocating spend as part of a connected TV strategy. The challenge of course comes with determining a single point of optimisation and attribution across the walled garden of Google, Instagram, and TikTok.

Whilst social platforms can look into their competitor's box of tricks to answer changes in user behaviour, advertisers create a new box of tricks. As innovation and duplication across social platforms become more frequent, therefore giving users more choice, advertisers are indeed forced to test more. Here the challenge comes not in what to test, but in how to measure it. We are forced to think outside of our usual box of tricks - which is already becoming more depleted by the quarter through cookie deprecation and changing privacy regulations, which lead us to question the robustness of measurement models we once trusted.

So is Media Mix Modelling the answer? Potentially, and especially if you ask Meta who released their MMM product “Robyn” to open source. Or, do we look to “dumb down” the measurement we put in place for these tests into the platforms themselves, allowing us quicker answers to move onto the next shiny new toy we’re presented with? Essentially it all comes down to what works for the advertiser based on their short and long-term priorities, available skill sets and the products that they have developed and own internally. Those who have more answers to these questions are the ones who will make the right decisions.

Oliver Williams, head of digital performance, Total Media


LinkedIn iconx

Your Privacy

We use cookies to give you the best online experience. Please let us know if you agree to all of these cookies.