There may be no more important business challenge in the future than data commoditization. Brands must begin setting data strategies to ensure customer data, especially the constant stream of behavioral data, will maintain business advantage over time. This is a difficult and hugely important problem to solve. Without data advantage in the future, is there any competitive advantage at all?
Data advantage can be achieved in both collection and processing. I define data advantage as unique understanding of your market and consumer needs/preferences within it – and using data activations to exploit your learning and keep you a step ahead of the competition. Your data moats will likely determine how well your business will fare in the future.
But what happens in the future when many others have the same or similar data sets? Your margin becomes an opportunity for everyone. In a market where most are using identical tools to collect and process data and building similar models, applications and data-driven services on top, how do you protect intelligence and marketplace advantage?
Differentiated Data Collection
Data collection is the foundation of the data value chain. Simply, data collection advantage occurs when data you can acquire can only be collected by you. This is a much more difficult task than many realize. Even with upcoming privacy regulations, data is collected from trackers that exist on your site, your app, your payment processor, your channel partners and distribution networks, just to name a few. Much of the customer data you collect now is to a large extent already a commodity.
Often there are multiple systems collecting similar information on your customers. Web browsers and the companies that operate them namely Google and Apple. Packets of data collected and inspected by Verizon, Comcast, AT&T or other ISPs. That is only the digital baseline. You have analytics, marketing channels, marketing services as well. And with the rise of IoT, data will soon be sent on your customers to the servers of all kinds of potential new competitors.
You can see the true effect of growing customer data collection commoditization best in advertising technology. Real-time Bidding (RTB) for display ad impressions also known as Behavioral Targeting (BT) is the most obvious example.
It’s a common occurrence in ad tech for companies to bid against each other using the same data. This is data that’s readily available about your customers and their behavior through vendor tags on your website and sold by data brokers to multiple parties. Over the last decade, data commoditization in advertising has drastically decreased the value of media (as measured in CPM or price per 1000 ad impressions). RTB and BT using customer behavioral data at the impression level became what many predicted, a “Race To the Bottom” on value.
Thankfully data commoditization in ad tech and much of ad tech itself is slowly disappearing with privacy initiatives. The winners left standing of course are the platforms because of their data moats. There are two ways Brands can get on offense against the commoditization of customer data collection.
One way is get strategic on “who, what, when, where and how” customer data is being collected. There are certainly meaningful data points of events, contexts and entities across the customer experience/journey you know impact your customers, your business and the market. There are probably others you are not aware of but brainstorming and listening to customers can suss out. All that can be leveraged into a first-party data collection strategy. One example is how customers self-segment based on browse behavior (btw an offline as well as online phenomenon). It is this “inside-baseball” on customer behavior and choice that brands need to strategize on collecting.
Another strategy rests in the exponential value creation of joining of data sets. You may have customer browse data that you can join with supply-chain data that can provide mutual value to your customers and your brand. Joining two (or more) different datasets together is one of the best ways to create data leverage. Leverage that can be used to keep the competition at bay.
Differentiated Data Processing
Processing is what you do with the data you collect. The end-result of a good data process are the predictive models and decisions derived from the collection and processing that are used in another area I’ve not addressed in this post, data activations. It’s important to understand that prediction is best served to improve the customer experience. What customers want, when and where they want it. What needs and intent are you able to fulfill that meets and exceeds their expectations? After all, customer experience is the deepest competitive moat of all.
Data processing will likely take a bit longer to begin commoditization than collection but when it comes it will come fast. There are always smarter people and always people that have experiences and insights that are unique and valuable. In the near term, while there is still a lack of people with data science and data engineering acumen, people are a competitive advantage. For a time, it will keep predictive marketing systems from being commoditized.
However, there are a finite number of models that exist and have proven to work for any dataset and market. Also, people will get smarter themselves and from the ability of machines to assist them. As we sit at the dawn of ‘AI as a Service’ it is a scary thought that this lends itself to even more commoditization over time. We’ve seen movie already with cloud services.
So, where do we get data processing advantage in the long-term especially using AI and Machine Learning (ML)? For processing to be a real differentiator we must come full-circle and highlight the advantages gleaned through strategic data collection. The ability to be accurate and improve the performance of the models over time starts with the data that goes into the models. For too many brands customer behavioral data collection is an afterthought turned over to a vendor. But we covered this in the above section!
One big defense against commoditization in the data processing lane is speed and accessibility of the data. We are beginning to see this now with marketing data latency. Many martech vendors can not provide data access for 24-48-hour windows. Other solutions have real-time streams of this same data type. It’s interesting to think how this data speed and access can get extrapolated across channels and systems for data advantage. It is an area to be strategic.
We have seen this previously in the rise of edge computing and co-locations in the cloud. We will see the same type progress in marketing technology. If you think about the newest hottest marketing technology, Customer Data Platform (CDP), CDPs core value props is centralizing data so it can become the new CRM.
However, in the future I believe you will see applications of AI/ML that sit at the edges of customer touchpoints. Things like web pages and shopping carts will have their own APIs to take advantage of the real-time capabilities present in data layers to make calculations and decisions in real-time at the customer touchpoint. A central data services model where decisions and calculation are pushed out at speeds that can’t match the speed of the consumer and the ability to create value will not win in the long-term.
Lots of brands and businesses are just getting started thinking about their customer data strategy but few are thinking about it with a competitive mindset. Most are thinking about it now as it relates to infrastructure. This is wrong. It is putting the data cart before the data horse. Customer data as a competitive advantage will only become more challenging and difficult to maintain in the future no matter your commodity infrastructure decision. A solid strategic foundation needs to be created for collection and processing.
In summary, you can begin to build your data moat taking these ideas into account:
- Data collection needs a strategy. It is the foundation of the data value chain. Every touchpoint your customer has with your business is an opportunity to collect data but the types and forms of data on customer behavior are valuable and important questions to answer. You need to format collection and schemas properly. Use cases are highly valuable here.
- Data collection increases in value by joining datasets together. To do so it needs a clear business purpose and a consumer value. The questions it will answer and the business cases it will serve need to be decided ahead of time. You may discover new nuggets of insight as you overlay datasets but mining for data gold is not a strategy.
- Your data processing will only be as good as the quality of your raw data. The customer data flywheel cannot spin without high quality collection, validation, enrichment and a constant flow of fresh high-fidelity data.
- Speed is a central component to data collection and processing advantage. If you are not real-time today, you are already behind. Increasingly, speed will be gained by moving calculations and decisions to the edge of customer touchpoints.
Last, I want to give a shout-out to my friend Jerry Neumann who inspired me to think about the subject of moats as they relate to customer data from his treatise A Taxonomy of Moats that should be required reading for anyone interested in the competitive nature of business.