Peter Thiel (ex CEO of PayPal) founded the company Palantir, a big data company, on the principle of hybrid human and computer big data systems. In the early days, PayPal had a whopping $10m a month fraud issue. To combat it they created cutting edge algorithms to detect the fraud… and they worked! Well… for a few hours. See, the issue was that no matter how good the algorithms were, they could always be beaten by the ingenuity of human fraudsters.
So, PayPal took a hybrid approach: algorithms did the bulk of the work identifying and flagging potential fraud and humans made the final decisions. This helped increase fraud detection and PayPal moved into profit the very next quarter compared to a previous $30m loss. This kind of hybrid approach is one that I'm very much signed up to on a day-to-day basis, I use it a lot for PPC.
I understand that algorithms and AI are amazing at crunching the numbers and able to process more in an hour than a team of the worlds best researchers could in an entire lifetime. I understand that its power grows each and every day. Its capabilities are improving rapidly, enabling it to plough more data through the likes of AI and Cloud processing.
So what's the issue?
Decision making and logic are being built into these programmes by fellow humans and it's now questionable to as whether it’s making the right decisions. And that's just it. I question any software that isn't white boxing its decision making, even the AdWords platform - when it is being vague, I cannot trust it knows better than I do.
I need to understand how and why it's reached that decision to raise my bids or adjust my modifiers. I also need to understand what results we get from that decision. After all, I'm the one accountable for performance and ultimately, I should know exactly what is going on in a PPC account.
Yet I still think with the right software, whether that is built in house, 3rd party or even AdWords, with the built in tools you can make smarter decisions and make more progress on an account. In fact, I'd advocate you going and finding a solution for your advertising platform and needs right now! Because, remember, computers are great at processing all that data for you, doing the heavy lifting, and allowing you to focus on what matters while also calling the shots, diving deeper and maintaining overall control.
We recently took over a PPC account that had been optimised by a custom-built piece of software for around a year by another agency. The software was making nearly all the decisions and it had run away with the account to the point where PPC performance was slowly degrading and the manager was struggling to identify what it was doing wrong. With some TLC and understanding of how the account was working, we were able to get the account back on track. This approach raised it out of slow decline and back onto the path of growth.
This hybrid approach is key to my management style. I have systems to alert me on poor performance, identify opportunities and to help me crunch more numbers. My role is to make final decisions, weave together any themes or patterns and use my tools to explore theories and gain insight into an account. I also keep detailed campaign notes and always know what’s going on in the account. By keeping my finger on the pulse, I can provide better insight and reports.
I believe in this hybrid approach but generally, the market is in the mind set that AI and programmatic advertising is the future. I don’t think either can be proved right or wrong at this moment in time, but I do know I’m making smarter, better decisions, reducing management time with mundane analysis, increasing PPC performance and adding value to my clients and my agency by using my preferred hybrid approach.