“If you want to make a ton of money you can help clients figure out their social creative. The agencies don’t want to touch it” Some choice words at my leaving drinks in 2016 after five and a bit years working for a big blue chip social network. Fast forward to today and the situation hasn’t really changed that much.
Creative on Facebook is still a big headache. And why do you think that is?
- Video is the most impactful format across Facebook’s services, but production remains too costly to justify for the majority of campaigns, particularly if you want good production values.
- Facebook ad servers makes rapid and often unintelligible decisions on creative optimisation within the first few hours of flight.
- Concepting to deployment takes weeks, with hundreds of emails, tedious Wetransfer links and confusion over ownership.
- Designers can’t keep up with the learnings to generate new imagery and often have limited insight into performance even if they wanted to.
- It’s hard for creative agencies to make a meaningful income off Facebook.
- Designers aren’t that excited at the prospect of making Facebook image posts anyway.
But the really big change between then and now is that machine learning has gone from buzzword adorning every stall at DMEXCO to an actionable suite of techniques which can enhance advertising for the better.
Not until very recently have we been able to make the most of image recognition, let alone starting to understand what elements of an image are responsible for driving the business impact you desired from campaigns.
Solutions to any question regarding social creative these days should make use of AI in some way. It’s the only manageable approach that can make sense of the volume of data now seen by consumers and make the advertising experience better. We’re now exposed to something like 1,500,000 ads per year…each! To convert this scenario into a musical analogy, that’s an awful lot of noise. So, converting that noise into sweet music which resonates with its audience is the role of the future of any solution.
Facebook sales teams speak to the virtues of iterative business practices with a test and learn approach. The theory here is rock solid, but if you implement this with people, it’s time-consuming and, let’s face it, a little jarring. Most people in the creative industries just aren’t that robotic. A quote from an industry expert still chimes: “We’re artists in a world of machines”. The industry needs a technology solution to implement business practices which are geared towards the dominant channels today.
To revisit a music analogy, if one is to ad-lib it means they perform with free rhythm and expression. Isn’t it really these principles which are being advocated by Facebook, just with a few guide rails in place?
Ad-Lib Founder Oli Marlow-Thomas agrees “we take existing assets and brand guidelines and use these as our guide rails. Then, through the axis of image recognition and Facebook performance data we automate the creation of new ads in real time.”
This approach means efficiencies in time to market, cost of outcome in media and the cost of creative production.
When, as with one of our case studies, the time given back to marketers represents a 90% time saving, isn’t it time to put the technology to work on a larger scale?
Harry Melsom, Product Manager - Facebook at Ad Lib.