With big data getting even bigger each year, marketers need to automate as much of the analytics process as possible, in order to keep things time-effective. This is the opinion of Vertical Leap’s Senior Data Scientist, Henry Carless, and he knows his stuff when it comes to handling data.
We thought it would be a good idea to ask him about how automation is allowing Vertical Leap’s PPC department make big things happen with data.
“We automate in order to be more efficient and creative with the campaigns. We use automation to take the work out of the analysis and gain insights very quickly.
“From the customer’s perspective, that means we’ve got more time to work on optimisation – making their account better. All of our automation tools are designed to give us actionable insights and take some of the legwork out of data preparation. We charge based on the amount of time we spend on an account; if we can use our time more effectively then they can get much better value for money, compared to an agency that spends the majority of their time on analysis.”
“One example would be identifying negative keywords.
“I’m looking at one account now where 10,006 different search terms triggered ads over the past few months. Advertisers normally focus on the most popular search terms because they can’t work through all ten thousand, but this means they miss the long-tail search terms that collectively have a real impact on performance.
“What we do is split those search terms into individual words and measure cumulative impressions, clicks and conversions for each word. This helps us find single negative words that appear in multiple long-tail search terms. By adding these words as phrase match negatives, we can negate a huge amount of long-tail traffic in a very short amount of time.
“So, this is an example of where we’ve taken the normal process of identifying search terms for negative keywords and by using automation we’ve turned it into a much more effective system that helps you analyse at scale, reviewing thousands of search queries in almost no time at all.”
“That’s a tough one. Maybe how we handle budget forecasting – that saves us a lot of time every day. Also, our quality score weighting system which helps us find the best opportunities to improve performance.”
“As with a lot of the things we’re doing at the moment we turn to machine learning for an answer. We use machine learning processes to quickly forecast spend for accounts across AdWords, Bing, Facebook – whichever platforms the customer is using.
“So, right now, I’ve got all my customers in one column and I can see whether they’re running AdWords, Bing or Facebook, the end-of-month spend forecasts for each, and whether I’m under or over spending.
“This allows me to say, for example, ‘right, I’m looking at a 31% overspend for this customer by the end of the month’. If the campaign is going well, I can phone the customer and say ‘look this is performing really well and there’s an opportunity to increase your budget by 31%’. So that immediately allows me to give actionable feedback to the customer on what they should be spending.
“If it’s not performing well, then I need to cut back the lesser performing areas in order to get back on track. This is key to optimisation because the first thing you need to know when you look at an account and decide what to do is whether you should be spending more or less.
“This is something a lot of businesses or agencies don’t put enough emphasis on. So they’ll look at the stats and maybe raise the bids for the keywords that are working and decrease them for the keywords that aren’t. But unless you know whether you’re fully utilising the budget or underutilising it, you could be doing completely the wrong thing.”
“Pretty much everything you’re doing in AdWords optimisation is to improve your Quality Score (QS). It takes time to identify opportunities to improve QS, because you might be talking about a hundred thousand keywords.”
“So what we do is we grab all of the keywords and look for ones with the highest impressions and a lower than average QS. These are the top opportunities for improving overall performance because they’ve got the most demand or the most visibility and a low QS. If you can make improvements here, this is where you’re going to get the most benefit.
“This can be invaluable data for a PPC manager: finding out where your biggest opportunities are to reduce cost-per-click and get more traffic for the same amount of money.”
“Actually, they get to see it for themselves through our automated reporting systems. For example, if we’re running a specific project for one customer and they want daily updates. If we weren’t using these systems, we would have to go in and manually collect that data on a daily basis and send that out to them – they’d end up with hundreds of spreadsheets sitting in their inbox. So instead, we create an online report where they can check their performance daily and it can be entirely customised for their goals.
“It does make a huge difference if you’re running a small business and your day-to-day sales are really important but you don’t know how to use AdWords. You don’t want to have to wait until the beginning of the next month to get a report that tells you how everything’s going. So, in that case, we can create a live, constantly up to date – up to the hour, pretty much – report that gives you all the stats at a glance.
So that gives you some ideas of how our PPC team automates the analytical side of things. This allows us to gain insights faster and build a much bigger picture of how things are working across accounts and our entire network of customers. But, above all, it gives us more time to focus on turning those insights into actionable PPC strategies that make a serious impact on account performances.
If so, just call us on 023 9283 0281 or fill in your details here and we’ll get back to you.
Michelle is the Marketing Manager at Vertical Leap.
We send a semi-regular newsletter on business and other related topics, with links to the latest stories from us and what we’re reading around the web.
Categories: Data Science, Machine Learning, PPC
Categories: Data & Analytics, Martech
Categories: Martech, SEO