The world is moving faster every day, and this is particularly true of customer facing-businesses. As people have more information at their fingertips (both to make purchasing decisions and to distract from them), it becomes increasingly difficult to catch their attention and understand their needs. It is common to compare old companies to dinosaurs, evoking the image of a lumbering Tyrannosaurus Rex, its tiny, useless arms flailing impotently as the world shifts around it. But what did the dinosaurs in was not their awkward size, lack of opposable thumbs, or even difficult-to-spell names (except for the pterodactyl – that’s just ridiculous). No, what drove the T-Rex and its brethren extinct was an inability to evolve with the situation. Most individual creatures on earth were doomed as soon as that giant meteor came along, but when the dust settled, the species that thrived were the ones whose members were diverse enough to find new niches, reproduce and further adapt, until they (we) became the new owners of the planet.
Your company is not a dinosaur (but your ideas may be)
Fortunately, your company is not like a prehistoric animal, with its DNA set from birth. It is like a whole species, with many related ideas constantly being born, fighting for survival, and recombining to form a new generation. For example, each time you run an advertisement, you learn a little bit, and that informs your next round of advertisements. Most advertising ideas “die,” either during brainstorming or after a tepid market reception, but as long as a few of them do well, your company overall gets to get to keep learning and growing.
We’ve now set the stage for the critical question: How quickly can your company’s pool of ideas evolve? If you have only one or two ideas at at time, and you stick with them until they have long since expired, you will probably go the way of the dinosaurs. However, if you have many smaller ideas, test them quickly and let the winners inform the next generation, you’ll find very rapid innovation and adaptation to anything the market throws at you.
How fast is fast enough?
That’s where analytics comes in. As Dr. Ryan Kirk (a consultant specializing in data driven decision making) points out, analytics can be useful across a range of time scales. Tracking key performance indicators (KPIs) on a weekly or monthly basis gives a strategic sense of how your organization is doing. That’s valuable for setting high-level decisions and is often considered the domain of Business Intelligence. At the other end of the spectrum, Data Science creates models that are much faster to run, sometimes to the point of being embedded in the user experience or the product itself.
There is a qualitative difference between reporting, even on a frequent basis, and the ability to move data so quickly that it actually informs conversations with the customer in real time. Anita Prinzie (product manager at customer experience management solutions company NGData) notes that moving from monthly reporting to continuously updated fast-moving customer profiles, enriched with fast-moving metrics enables real-time decision making, allows companies to respond to customer interactions in a timely, personalized, and relevant way. This is a critical differentiator in the quest to improve customer experience.
Bad mistakes, worse mistakes
There will be growing pains. If you’re selling T-shirts, and you accidentally send a few in the wrong size, that’s not a huge deal. But if you are selling food products and accidentally suggest something that your user is allergic to (even if they see the warning before hitting “buy”), you will have lost a lot of trust. It’s risky to make your computer a salesperson, but it’s also worth remembering that it will get better with each iteration.
The chess-playing computers lost to top-rated humans for years, but developers ultimately ironed out the kinks to consistently achieve victory. Likewise, your algorithms will need to make some mistakes at first. Eventually, the investment in time, money, and placating a few confused customers will pay off in a better experience for everybody.
It’s very hard work! Having systems that are part of customer interactions increases both the number of inputs and the number of outputs that you need to consider.
- How does the customer input data into your system?
- Where does the output go, and does anyone interpret it before it goes in front of the customer?
- If it’s going in front of the customer immediately, how tolerant can you be of funny looking output?
- What will you display if the data is incomplete?
- Are you inadvertently exposing someone’s personal information? Perhaps, for example, there is a small sample size, and you’re calculating the “average” of just one person’s data?
- What is the value of making the optimal decision versus a first guess?
- Is there a culture of cooperation between your data engineers and your front-end engineers?
- Are you sure the data is accurate?
- Do people trust the data in the first place? (Note: This is a different question than “Is the data accurate?”)
Mark Harrington of Clutch, a leading Consumer Management provider, points out that making strategic decisions that are devoid of certain channel dimensions is risky, but often necessary because of the complexity of integration. This is why you need a central customer marketing platform that serves as a “hub” that can integrate with all of these channels to ingest each’s data to synthesize it into a complete customer view. Once you have this and start increasing your “Customer IQ” you can begin motivating your customers and earning genuine loyalty from them.
Business Intelligence and Big Data
None of this is to say that the Business Intelligence side of things is not also important and rapidly evolving. Big data technologies bring several different benefits, often described in terms of “The Three Vs.”

Credit: https://www.flickr.com/photos/carbonquilt
The first is Velocity. That’s what we discussed above, the ability to harness lots of information in the blink of an eye to deliver real-time, data-backed results (for example, twitter data, as simulated in this post). It also includes Volume, which is the ability to use data that is extremely detailed across a wide range of times spans. It’s amazing, but until now many organizations stored much more data than they could effectively analyze. With big data technologies, they can finally crack open the vault that they have been guarding for so long and see trends over the course of years or more.
The final V is Variety, which means that you can take advantage of data that you hadn’t even really thought of before. A common example is free text, which is awkward to process in SQL, but which can be handled in more appropriate languages with MapReduce technologies. Another example is image data. Your organization may have thousands of images stored for operational use, but you never really thought that analyzing them on a measurable scale would be feasible. Audio data is the same; it’s hugely computationally intensive to distill into insights but is now accessible to any company. Ask around – it’s likely that your organization has been storing these assets for operational reasons, never imagining that they could be parsed into useful features, aggregated, and used to drive decisions.
The opportunity is huge
In addition to improved storage and processing technologies, there are a wide variety of algorithms available for plug-and-play use at a low cost. IBM Watson, the same computer that beat the best humans at Jeopardy, has an API available for a wide range of tasks at a low cost and a mercifully shallow learning curve. Don’t worry about whether this makes you a ‘fake data scientist‘ – this and other technologies allow organizations to get a sense of what they can do with their data without hiring a team of PhDs. With the cost of small-scale experimentation being so low, the question shifts from “How do we fit this into the budget?” to “Am I excited enough to spend a few hours on this?”, which is where the real experimentation and innovation begins.