I’m usually prompted to write after a fun customer call, and today is no different. Chatting with some folks about challenges delivering data to the business, I asked “What’s your biggest pain point? Not knowing what data you need, what questions the business is asking, what the next request will be?” Their answer was a resounding “yes.” While some might see a frustrating problem, I was really happy for the team – it meant they were shifting from the world of business intelligence into analytics.
If you want to understand what’s happened in your business, hire a good BI team. If you want to create deeper insights into your consumers, get a good analytics team. I segment business questions by the time it takes to get to a relevant answer, as well as the value to the company (tactical or strategic). A simple 2x2 grid visualization with examples is here:
One of the common issues that I've seen with many clients is that most analysts and operational teams live in the top left box. A question arises, analysts request information from (often) the BI team, wait for response, review the info, find a nuance to dig into, and start the cycle of request-wait-review all over again. By the time you have an answer, the question has been forgotten.
Similarly, I think of BI and Analytics along a continuum: BI tells you what’s happened, Behavioral Analytics tells you the who and the why, Predictive Analytics tells you what’s going to happen next and who is going to do it. That same 2x2 now looks like this:
The magic is in realizing that these aren’t discrete steps or areas – you can and should move up and down that arrow, and hopefully you will be driving analytics closer to the business user / internal consumer. Many companies split BI from analytics and data science, where they're seen as either irrelevant to each other or worse, in a competitive situation.
One of the first places that we miss the opportunity to move smoothly across the timeline is with how we feed data to business analysts. A day in the life of a business analyst for a call center involves reviewing lots of KPIs. Each day, check if a metric broke trend, how much did it break and enter a ticket for their BI team to pull some data so they can try to figure out what’s going on. If they’re lucky, they get their data a few days later, begin to manipulate it in Excel, put together a presentation and can answer on Thursday of next week what happened on Monday of this week.
What’s the solution?
Start using your analytics both in a behavioral model as well as longitudinally (that’s a two buck word for “across time”). Take the plunge and supplement your current demographic and product models with insights based on expressed actions online, offline, across your CRM and IOT systems and social media. As you combine deeper insights into the experience your consumers are having with you, you can better predict the trends in contacts, comments and sales. If your IOT systems tell you that a consumer just had a product failure, you can expect a contact (call, email, web visit). If you aggregate those parallel problems, you can now tell if you’re breaking out of trend relative to past situations.
Imagine that earlier business analyst comes in on Monday, pulls up the daily reports and sees three KPIs highlighted for having broken outside predicted tolerances. As they click down, they see the primary issues that drove call volume (“failures impacting products x,y,z”). With clustering tools, the analyst can evaluate if there are similarities (firmware, geography, weather systems, etc) across those problems and initiate a high-value and timely response to the problems.
Accelerating time to answers is one of the first and most essential ways that you can make analytics relevant to your business. You are creating actionable analysis, your analysts have more challenging and satisfying work, and you can respond to changes in the market or emergent threats more quickly.
What does this mean to your business:
- Faster, more individualized responses to your customers = happier customers, higher loyalty, higher retention.
- More engaged workforce through challenging work = higher job satisfaction, lower turnover, higher value per employee to the business.
- Quicker response to emerging problems / threats = address issues before they become risks, early identification leads to both cost savings and protection of reputation.
What happens next?
As you start down this journey of accelerating answers, you will come to the next step beyond this: once you have your analysts doing high value work, how do you begin to automate the response to an issue, so that it no longer requires human intervention? That's the topic of a future post - enhancing the experience through educated guesses.