How will Digital Marketing adapt with rising AI automation and machine learning?
We have heard enough of jobs that will be taken away by AI. I’m not worried about that. Not because I don’t believe in machine not capable of doing today’s jobs but the fact that this has always been happening since a long time. The world will adapt and find new jobs. There was a point when a postman for personal letters was a real job. We don’t see them nowadays. Anyway, that’s a story for another day. I want to pen my thoughts on how the digital marketing industry is going to change with Automation.
Okay, if you are new to digital advertising, I’ll break it down to the different broad parts where marketers play a major role.
- Targeting and Creatives
- Campaign Management
- Tracking and attribution
- Reporting and Analysis
Targeting and Creatives
The combination of targeting and creatives play the most important part in advertising. Every channel will offer their own form of targeting, for example Facebook offers location-based targeting whereas google SEM may not. Targeting is best when your audience is narrow as possible. To find that you need to use different small combinations and try them out to find the best one. This method is expensive as narrow audiences will also cost you more. So companies like Albert.Ai are trying to fix. However these platforms need to be far more mature and to actually take it away from humans. A silent competitor to these companies is Facebook and Google itself since they control the data and do not provide enough pre-click data for others to optimize. However sooner or later, we’ll see a fully automated targeting algorithm which will take away to pain of trying out combinations. We’ll still need to provide the initial broad audience.
As far making creatives, I don’t see a fully automated system anytime soon. Creatives are the souls of every marketing campaign and a robot cannot simply replace a creative designer. But, there are already other forms of automation which have started playing with combinations of ad copies. For example Smartly.io started a feature where they automatically add texts from the feed to images for DPA in Facebook which makes it easier for the marketing team to try out combinations without going to the design team.
Campaign Management is where most Ad tech companies are trying to automate things. Because it is easier and also at the same time the human effort required is too high and too mundane. Most companies have started off with rule-based engines like Smartly, ROI Hunter etc. Other companies like Nanigans, Albert, Media Math, TradeDesk have invested more in data science. Currently the former is still highly accepted by many companies because some marketers are reluctant to go for tools which do not give them control over their ads. Others are just scared it will take their jobs away. But again it is inevitable that Campaign Management will go away few years down the line.
Tracking and Attribution
Tracking and Attribution plays a huge role in the performance marketing industry. Simply because when it comes to numbers, there is mismatch everywhere. The truth is only your CRM data. The reason behind this is attribution. In short, suppose a guy lands on a site from Channel A and then doesn’t buy the product. He lands again from Channel B and buys the product this time. Both channels A and B will claim to the transaction. However, the result was a process of multi-touch attribution. There are already many companies who are into the ad attribution business like Appsflyer, Google Analytics, Convertro who takes care of this. AI and ML are playing a significant role in improving the attribution process but a marketer has no role to play here so not going into details.
Reporting and Analysis
The last step of the cycle is reporting. It’s said that 10% of the campaigns produce 90% of the results. The goal is to find these 10% campaigns. Reporting can be a tedious job with a lot of data processing and analysis something which has become out of hands for marketers and has invented a new job title — Marketing Analysts. The short process is to take pre click data and join it with post click data and make splits and cuts to find those 10% campaigns. However, this is very naive. There’s a joke which goes around Water with Vodka destroys your liver, Water with Whiskey destroys your live, Water with Brandy destroys your liver so water is the reason. This is very true in analytics space. Because analysts don’t have the time and resources to go deep inside. If they went deep they would have found that alcohol is the reason not water. Which is why there are new companies like Adwyze, Datorama, Nuggit etc which have raised in hundreds of millions to solve this problem. But yes it is also inevitable that 90% of this job will be automated in future.
Though this article has been mostly on performance marketing, brand marketing will have its own form of automation. But before brand marketing automation happens at a very strong level, there needs to be more work done towards attribution for brand marketing. After attribution becomes perfect, brand marketing will become similar to performance marketing. But it is certain that in sometime in future 60–80% of today’s marketing jobs will be automated. Thanks for reading this article. There’s a video interview from the founder of Machine Zone which I like a lot. Have a look into it if you would like to learn more about.