— June 8, 2018
This is a case study about how our data scientists are writing automation scripts to free our PPC specialists from hours of manual work, enabling them to focus much more on strategic and creative tasks.
Optimizing a Google AdWords campaign used to be simple: Some keywords are more valuable than others, so we would bid higher on those that are more valuable and drive budget into the best performing keywords.
AdWords is now over 17 years old, and a lot has changed since its inception. We no longer target keywords, we target people. In 2018 we have much more control over who sees our ad – we can tell a lot about our audience and use this to tailor our bids, helping distinguish between high and low value users even if they’re searching for the same keywords.
In the example below, we start with a £1 bid, adjust it according to the various factors and end up with a bid of 97p.
Let’s take a look at how long it takes to do make a single mobile adjustment bid. These are the steps that a PPC specialist would need to go through:
- Check there’s enough data
- Check the result is statistically significant
- Calculate the bid modifier
- Set the bid modifier
This takes approximately 20 seconds per device category, per campaign. If you optimize for two device categories across 1,500 campaigns, this equates to two days work per month.
Now, if we multiply this by all the other modifiers that exist e.g. 20 cities, six age brackets and so on, the time required to optimize all of these bid adjustments is approximately 500 hours! This is unfeasible and unrealistic given all the other tasks that PPC specialists need to do on a day-to-day basis – there simply aren’t enough hours in the day to do it properly.
Because of this, most agencies estimate their bid adjustments. In our example below, you can see again on the top row our bid of 97p and then on the row below, an example of where adjustments have been rounded to the nearest five or 10 pence. As a result, the final bid is only 72p instead of the optimum bid of 97p – this would mean missing out on valuable traffic!
So how do we solve this problem?
In a word – automation. We have data scientists in-house that create scripts and systems to do all of this manual work for us, freeing up time for strategic and creative tasks which can’t be automated.
Once this has been set up, the time it takes for our PPC specialists to perfectly optimize the bid adjustments is 20 seconds per category (device, location, gender), per account.
Referring to the example below, this equates to 3.3 hours – work which would have taken 500 hours to do manually with the same level of accuracy.
This is a great example of how automation is helping us do a much more thorough and effective job for our customers, where less of their budget is being wasted on manual work and more is being spent on helping them get the competitive edge.