Simple:
Transparency = Knowledge transfer
Knowledge transfer = Market growth
Anyone can have a million dollar idea, but only a few can make it a reality. This is why we went above and beyond the conceptual framework in order to spark an open conversation about this in the market.
Yes and no. We can, and will, automate a lot of the heavy lifting processes, but the Search specialist’s job will inevitably change drastically over the next few years. As of now, even Google is trying to kill the Search game that made it rich. Search specialists were a unique breed, and we believe they’ll survive and evolve as Inbound Algorithm coaches.
Adviso is currently working on this with its Age of Ultron project. It is a project that we started developing in 2017 and believe we’ll have ready in 2018.
For those who don’t know the origin behind the ‘’Ultron’’ name, here is a bit of Marvel Cinematic Universe insight: Ultron was AI invented by Tony Stark in the Avengers stories. It was created with the intention of helping protect Earth from various future threats, but ended up identifying Earth’s inhabitants as its own primary threat.
We named our project accordingly because, while we are trying to help our specialists in their day-to-day work, there is also a possibility that we will forever destroy our relations with them. We must tread carefully because we believe in human inventiveness.
First, let’s say that a specialist sees every possible keyword as a slot machine. With this logic, showing your ad may give you a payout, or may not. A good machine will probably not payout on the first try, but may be very profitable in the long run. How does one correctly optimize his test in such an unsure environment?
Normally, search specialists will proceed by instinct first, and by numbers second.
We are currently working on replicating that with a K-class algorithm, which we’ll present to you in a future post. Essentially, the script looks at the campaign’s subject, then it draws a keyword graph and starts looking for new keywords suggestions based on various tools. Then, the fun part begins.
We simulate the impact that each and every new keyword would have if they were added to the account. Then, we automatically classify them in the actual structure by either proposing them in an existing ad group or propose a completely new group if required.
The idea is to exploit the Multi Armed Bandit optimization technique.
You can also use the algorithm to generate a completely new campaign from scratch that you can then you run afterwards. You then layer other tools on top of it, like Go play in Traffic, and use the Blindfold to automatically give more credentials to the best performing search terms. Then, you continuously research new opportunities with the Multi-Armed-Bandit optimization technique using the keyword classifying script to keep an updated feed of opportunities and potential payout. Which renders the “exploration” part less tricky by letting the AI manage the global budget and goal, and ask it to continuously test new things while also continuously delivering better performing results.
In this scenario, the specialist’s job is now to look at the account’s growth, make sure it is aligned, create powerful messaging and focus on ensuring the best user experience possible : from the queries to the message, to the landing page, to the service delivered.
With this, you have a fully automated account that follows a path that a specialist would normally take by instinct. He can now focus the bulk of his efforts on understanding your audiences and optimizing your messaging techniques.
This is being a coach. You point in the right direction and use the tools to achieve great success. We believe that this is where the future of Search specialists lies.
The following example from Randal Munroe of XKCD tells it all.
From looking at the image, try to answer this : What happened here?
Anyone who has ever been in contact with a child will give you this answer without hesitation; the kid will say a) or b), but the truth is c). One of our strengths, as humans, is our ability to grasp heuristic data and do post-performance analysis. It’s our ability to look at a result and figure out how it happened by asking the right questions. This is where you should invest in your team training over the coming year: Learning to ask the correct questions.
To support this conclusion, bear in mind that Google Analytics, Watson and many more are developing tools that will replace Excel and manual data mining. You’ll simply send the raw data to the interface, ask a question and ponder the result. Watson analytics licences cost about 30$ per month.
The winners of the future will be those who will be able to drive the AIs to their limit rather than being driven by them.