Making the Big Advertising Algorithms Work for You
Six years ago, I made a decision to invest in marketing and in myself. Three years ago, I completed the journey that decision enabled and earned my Master’s degree in statistical analytics at the University of Chicago.
As a marketer, I believed then, as I do now, that strong mathematics was going to be the differentiator between good and great marketing, especially in the digital space. To justify this assessment, I want to share a question I get often from my clients: “If everyone uses the same algorithms that Google, Facebook, and Microsoft provide, how can we differentiate to beat our competitors in this space?” Certainly a great question, let's take a look.
About Algorithms
Having spent countless sleepless nights coding, testing, and improving algorithms, I can confidently say that the code itself represents well less than half of the overall likelihood for success when applied to business scenarios. To simplify: algorithms (as with many things in life) are susceptible to ‘garbage in, garbage out’ effectiveness. An algorithm has three core components that all must align for applied success:
An organized, rich, and transformed data set
An algorithmic model
An optimization function
It is not my intention to write a math-heavy blog that is only valuable to the most math-centric readers, so we will spend most of our time focusing on the data and optimization opportunities and let the smart people at Google continue to work on their model (though we will address that at the end).
The Role of Your Data
Let's start with data, the secret sauce for success.
In this case, data will have 2 core components:
‘What’ data is included (and excluded)
‘How’ the data is transformed and fed into the model
Choosing the Data to Power Your Algorithms
More data is not often the best solution. In fact, most good model building will first address the relevance and significance of input data to model success and remove inputs with high variance, which are typically unreliable or not valuable to the end result. In search, this data comes in the form of the keywords, keyword strategy, organization of keywords (often into campaigns), and additional behavioral and ‘results’ data based on the engagement of those keywords. It's easy to oversimplify this process, especially if you ‘over apply’ human opinion. At Rise, we are constantly identifying keywords that may not be the most intuitive additions to campaigns but do deliver measurable business impact. A smart digital marketer will allow this discovery process to play out, let the data lead the way, but all too often we find ourselves limiting our own digital horizons due to ‘gut feeling’ or ‘experience.’
Organizing Your Data to Work Properly
The data needs to be organized and formatted in the right way to enable maximized results. This can be a painstaking process but is absolutely unavoidable. Each data point should be reliable and mirrored in format to other figures within the same metric.
Granularity of build is often the “secret sauce” many agencies rely on to beat their competitors. I refer to this build often as the fingerprint of an agency. Each business will benefit differently from the nuance of search structure and it's important to understand the outcomes the business needs so the most optimized search account can be deployed.
Building Your Optimization Model
We now have strong data, a world class structure, and of course a platform-provided algorithm with which to work, so what now? The next decision can be critical and is often under-appreciated: the optimization model.
At its simplest, the algorithm needs to understand toward what (specifically) you want to optimize. This is both a business question and a data/technology challenge. To this complex question, less-sophisticated clients often answer simply “yes,” thinking “of course I want to optimize my campaign throughout its run!”
Why doesn’t “yes” work? The question isn’t merely “Should we optimize?” It’s multi-part. Some examples:
“What would you prefer to focus on, efficiency or volume?”
“Is it important to spend budget in full?”
“Do we need to hit a specific cost to ensure business success?”
What goals are the client marketers focused on, and how much pressure is there to achieve said goals? This challenge is compounded by the added challenge of delivering on the optimization once the decision is made. For example, if we optimize for efficiency, how then do we quantify the “value” of said efficiency? Simple efficiency metrics such as Cost per Conversion or CPC are marketing-centric and not truly business-centric.
It is essential to apply key business nuances to these values, even though they are more complex and often enabled through calculation and technology. For example, utilizing your own product margins to reward sale of high margin items, or perhaps factoring in lifetime value for B2B. You could take that one level deeper in complexity and apply calculations for incrementality, enabling less reliance on branded terms and opening up non-brand opportunities.
Using Algorithms to Beat Competition
Mathematics in digital marketing was once a tremendous differentiator between the organizations that used it, and the ones that didn’t. However, over the last few years, we have seen increasing parity in the digital space and now excellence in execution is the hallmark of real differentiation. If nearly all digital marketers use Google, and hence the algorithm and tools within, what can a professional marketer do to be the best?
The Google platform can ingest your properly-formatted, business-centric calculated values and in fact optimize towards them, creating a very customized and focused application of the “same algorithm everyone uses.” Though we will all use the same Google algorithm, the inputs we enable it to ingest combined with the custom calculated optimization method we aim it toward create a totally unique and expertly applied approach which can easily differentiate you in the crowded digital space, yielding a competitive advantage to build on.
Rise can help. Our award-winning team + technology approach to digital media combines granular account optimizations with tech-enabled intelligence to get our clients real-time insights on their biggest opportunities for growth. Reach out today to talk with our team.