Original article can be found here (source): Artificial Intelligence on Medium
4 Overlooked Questions You Need to Ask Before Creating a Data Strategy
The world is awash with data, of which 2.5 quintillion bytes are created every single day, and globally, 90 percent of all data was generated in the past two years.
This represents a huge opportunity for marketers, but it also brings challenges. Obtaining data is the easy part — the question now is how do you ensure that data is valuable and useful? How do you make sure it serves your business best by helping you formulate effective marketing campaigns?
Here are four essential questions that are usually overlooked, but should be asked before creating your data strategy, in order to make sure that your data is valuable and serves your business’ goals.
Q1: How Recent Is Your Data?
Data recency is a key metric when building a data strategy. The more recent the data is, the more valuable it is, as it more closely reflects a consumer’s ever-changing behavior.
Using analytics tools, you can assign your customers a recency score based on their recent visit to your website and the interval of a purchase they recently made. This will help you filter out new users, for example, if that would be helpful for your marketing strategy.
This will help you target your marketing spend more effectively. If you are promoting your business daily to your customers, but most of them only visit once a month, that would be a huge waste of your marketing budget. That segment of the budget would be better spent targeting those customers who visit more frequently, or crafting a bespoke marketing strategy aimed at converting these less frequent visitors into more habitual customers.
Hence it is wise to refer to a recency score before launching any new campaigns. That way, you will avoid bombarding your customers with the same old marketing materials, and instead segment your audience into more highly targetable tiers.
Q2: How Noisy Is Your Data?
Not all data is created equal. Sometimes you can be faced with ‘noisy’ data — this is where some values or labels are wrong to some small degree. A few percentage points might not seem like a lot, but it is enough to corrupt or distort the data so it tells a very different story to the one told by the ‘true’ data. This can lead you to create a misguided marketing strategy.
In these instances, you need to look closely at other variables to see if you can leverage them to remove the noise. Another way of ironing out these wrinkles is to aggregate the data into a greater data set.
For example, in touchpoint data, one particular cookie might receive 100 display impressions in a row from the same website. In this instance, you must decide whether this is a blip and so should be treated as just one impression, or if there were actually 100 impressions in such a short space of time. It is worth bearing in mind such potential outliers, as — if undetected — they can wreak havoc when it comes time to analyze the data.
Q3: How Diverse Is Your Data?
The more data sources you can use, and the more diverse those sources, the better. That is because data pulled from a wide range of diverse sources — as long as they are all relevant — will give you a more rounded picture of your customers’ habits and behaviors.
Using a single data source is a road to nowhere, as your data will be very limited. You can add extra dimensionality to your data by adding more sources, such as purchase history, customer profile information, search behavior (both on your site and on the wider web), and campaign data. This will let you analyze your marketing data in much greater detail, giving you more insights upon which to act.
Q4: How Fast Can You Feed New Data to Your Machine Learning Pipeline?
Machine learning is by far the quickest and most effective way of selecting useful data sets. However, machine learning — and any AI model in general — is only as good as the data you feed it. So, you should make sure you have satisfactory answers to the previous three questions before feeding data to your machine learning pipeline.
Then speed is of the essence. Otherwise, you create a bottleneck in the data lake, which will compromise the recency of the data, and hence its efficacy. It will be critical how quickly you can feed the data to your machine learning pipeline in order to test it out.
Flexibility and scalability (up as well as down) will also be crucial considerations when choosing a machine learning model. So make sure you pick one that will serve your business well as it evolves in the years to come.
Data is invaluable in creating marketing strategies, but you need the right data, treated in the right way. By asking these four questions, you will drill down to the really useful data, generating the kind of actionable insights that are essential to develop an effective marketing strategy.