After winning Google’s first Performance Agency of the Year in 2017, Columbus was rewarded with an exclusive agency Google summit. 5 Columbians were invited to Google HQ in San Francisco and made the most of the money can’t buy experience, absorbing information to share with our clients.
Columbus was lucky enough to be exposed to an array amazing speakers and presentations in a jam packed two-day summit. There was an array of content covered, ranging from product roadmaps, data transformation, Growth Labs to Human Resources and people development. Now that we have returned to earth, we want to clue our clients in on what is coming from Google and how to gain an advantage with early implementation.
Value Optimisation will be the first part of a 3 part series by Columbus’ National Head of Performance, Mark Duffy.
Value optimisation. How to capitalise on machine learning and harness the true power of data.
Not too long ago, Performance marketers faced huge limitations in optimising for value outcomes for clients. Placing a conversion pixel on a website was simply deemed a success. It meant marketers could track conversions that fired a pixel when a consumer reached a ‘thank you’ page. Bid strategies were habitually set up using Cost per Acquisition targets that were identical for all customers, regardless of the value of the customer. There were a lot of unknowns; did the consumer book a more expensive (and more profitable) long distance flight? Did the customer add-on extras when buying their health insurance? Was the cost per lead for the test drive for a BMW 5 series M or a BMW 2 series?
Whilst this approach has its flaws, it has clearly enabled Performance agencies to deliver value to clients and continue to play a key role in the consumer journey. However, as with everything in life, there will always be a better way. From a Performance Marketing perspective, that better way is to optimise campaigns for increasingly profitable and higher value outcomes, delivering more business value to clients.
What has changed?
There are two key drivers of this change that have allowed us to shift value. As obvious as it may sound, these key drivers are data and technology themselves. What isn’t so obvious though, is how exactly the two are merging together in ecosystems to identify ‘value’ leveraging your existing customer data with machine learning technology and find more valuable customers across the wealth of Google properties, in particular, Search, YouTube and the Display network.
Feeding the machine.
From a data perspective, businesses can leverage their own first-party data and segment out customers who are deemed of higher value. This higher value may be determined by, the frequency of purchase, higher order value, low claim rates etc. When this high-value pattern emerges, these value segments can be pushed through to analytics platform, such as Google Analytics 360, wherein they can be further classified into logical segments (high lifetime value, high average order value etc.) By passing these segments into Adwords and implementing Smart Bidding programs, segmentation and machine learning are brought together. This machine learning stack is, therefore, able to identify patterns from the 20M+ digital signals that match those existing high-value customers and then targets those lookalike segments. Once identified machine learning prioritises bid strategies to high-value audiences instead of basic standard conversions.
This was one of the most inspiring and thought provoking sessions from the summit. There is a clear opportunity that this evolution represents for our clients. There is no denying that the set up for machine learning will involve a significant amount of time and commitment from a team. Google Analytics 360, Google Deep CRM and Adwords will need to be utilised and working in unison. However, this is long term solution that has significant long term benefits. Being able to continuously find and acquire high-value audiences in a digital ecosystem, who spend more money and have higher lifetime value, is priceless. We are no longer applying blanket CPA targets to acquire all customers regardless of their value. Instead, we are optimising for value using machine learning and harnessing the truest power of data.