• Analytics
      Consulting

      Are we a
      Good Fit?

  • Pricing
  • Contact Us

Blog

AdWords | Man vs Machine Learning 2018

If you are a nerd like me and read about machine learning then you’ve most likely heard about how algorithms can beat humans in chess.

AdWords success shares many similarities with the game of chess.

In AdWords, you are competing for spots on Google’s search engine results page, or other placement options, and instead of trying to capture your opponent’s King, you’re trying to generate a conversion.

So, if humans can create computer programs that consistently beat the world’s top chess players, is there such thing for AdWords?

Well, kind of.

I say kind of because there are really two different types of moves in AdWords.

  • moves that require a human
  • moves that should be made by machine learning

Moves that require a human are; initial setup, conversion tracking, creating your product, your brand, your messaging, your fulfilment process, your customer service process, etc. These are like your actual chess pieces for your business.

Once you create your chess pieces and place them on your board (AdWords account) then that’s when machine learning needs to take over and calculate what your best move is in real time.

Moves that should be made by machine learning primarily involve bid management, keyword optimization, ad rotation, time of day, day of week, age, gender, context, geolocation, device, past browsing history, and a handful of other things that can impact success.

Just like in chess, the goal in AdWords is not to make as many moves as possible, in fact the goal is to make as few human moves as possible to beat your opponent (winning = consistently generate conversions for your business at your performance target).

In case you have never tried to play the game yourself ( generate profit using AdWords), let’s just say that it used to be a whole lot more time consuming than it is today.

Pre-machine learning, the day-to-day responsibilities of human controlled bid management involved spending countless hours crawling through the accounts historical data with pivot tables (the same performance data that the machine learning algorithms now automatically analyze for you) with the goal of making bid adjustments that are predicted to make future improvements

Let’s just put it this way. similarly to how no average, or even above-average chess player could ever expect to beat a machine in chess, a human AdWords manager should never be expected to make better bid optimizations than a machine learning algorithm. It’s literally impossible.

*Unless you actually do have one of the smartest mathematicians in the world spending 24 hours per day analyzing the most recent data, but then you would have to pay them for their human time which can be extremely costly and even end up costing more than your ad spend itself.

So where am I going with this?

You most likely won’t magically start generating thousands of sales by creating a campaign and letting Google’s or Facebook’s bidding strategies attempts to hit your performance target. You have to make sure all of your chess pieces are in the right place or else you’ll get beat and can end up wasting a lot of money.

If your business has live AdWords campaigns right now, then you need to ask yourself;

  • Are all of my chess pieces in the right spot on the board?
  • Are my moves being made by a human or machine (i.e. campaigns being optimized in real time by automated bidding strategies)?
  • Are my campaigns generating profit or am I burning money?
  • Is my performance trending up or down over the past 30 days?

If there’s one thing that I can tell you for absolute certain after looking at thousands of live AdWords accounts, it’s that historical performance will most likely predict future performance. If your AdWords account is not generating profit today, then it most likely will not tomorrow either.

What I preach here is not a magic bullet or anything fantastical.

Google and Facebook have invested billions into developing their machine learning applications. These machine learning algorithms are readily available for any advertiser who uses AdWords or Facebook.

The challenge is, how do you set up your chess board? Because I can also tell you that if your chess pieces are not setup correctly then no matter who is managing your bids, you will most likely fail.

If you learned something from this post and have an active AdWords account that needs to be optimized to generate more profit, then give us a call.

We not only help with bid automation, but we are expert Consultants on all aspects of digital advertising and can help you make sure all of your chess pieces are in the right place to set you up for success.

We look forward to optimizing your AdWords account!

Share:

Facebook
Twitter
LinkedIn
Email

Related Articles

Get In Touch

GA4 Basic

Transfer All Universal Analytics Tracking to GA4

Anything you are tracking now in Universal Analytics we will make sure you are also tracking in GA4.

Set Up Conversions in GA4

You likely have a list of important conversion actions that inform your core KPI’s, we’ll make sure those are all set up.

Set Up Audiences in GA4

GA4 has a slightly different way of grouping data that relies heavily on proper audience creation, we’ll make sure your audiences are set up in a way that makes sense for your business objectives.

Integrate Big Query to Preserve Data

If you don’t set up the Big Query connection you’ll lose your historical Universal Analytics data, we’ll make sure this integration is done correctly.