WHAT IF?
You Could Plan Future KPIs

COT WP
“If only we could predict the future,” goes the saying. Throughout time, people have been fascinated by the chance that there might be ways to find some advantage, whether in love or money or happiness, via the well-worn crystal ball. We’re not there yet, but the analytics potential of the IoT is bringing us closer.

For business managers, the new blessing of more and faster data also creates new pressures to get the analysis, tools and methods right for mining that information, and to do it better and faster than the competition.

Here’s a look at some of the ways that the heart of your analytics, your KPIs, are likely to change, and some ways they’re likely to stay the same.

How KPIs Will Change
Major changes in two main areas: There will be whole new KPI categories, and KPIs will become more forward-looking as IoT analytics mature.

One example of a new category that wouldn’t have been considered a few years ago is in the realm of social advancement. Investigations of entire communities will become commonplace as researchers look for ways to foster compatibility with innovations in “smart-city” technologies. For instance, the study of community KPIs will measure the degrees of acceptance of everything from autonomous automobiles to reshaped neighborhoods and work places.

Improvements in forward-looking KPIs will come from the combinations of more granular data and more capable modeling techniques that make use of advancing machine-learning techniques. The difference: so-called lagging indicators, which tell what has happened, will be regularly replaced by leading indicators that will predict what is likely to happen. An example would be in complex manufacturing processes such as semiconductor fabrication, where predictive models will lessen the reliance on expensive and time-consuming post-production tests to improve manufacturing yields.

How KPIs Won’t Change
Best-practice principles of designing and extrapolating meaning from KPIs will remain as important as ever, and will challenge product managers to maintain focus on the most meaningful aspects of the higher volumes of data they’ll be seeing. For instance, it won’t be any more necessary to total up sales numbers – something that should be left to the sales manager – than it is today. But it will continue to be important to determine which features are being used by which customers or why customers are abandoning a product in favor of the competition.

How You Can Prepare
In a word, stick to the basics of good analysis and smart development.

Good analysis means asking the “Why” question once you’ve determined your most critical KPIs for a given product, or a given strategy. The concept of asking Why, sometimes again and again, was brought into popularity by the founder of Toyota Motor Corporation in 1958, Sakichi Toyoda (the spelling is correct), who said “Ask ‘Why’ five times about every matter.” Asking Why – often more than just five times – is the basis of root cause analysis, the well-regarded process of end-to-end analysis used in manufacturing and other industries for investigating malfunctions. It’s good to use it in designing KPIs, too, both now and in the future.

As for building those future models, don’t be afraid to rely on best-practice development techniques that are known today. Especially in building time-series models, which are central to capturing and predicting trends, opt for low-code, reusable, building-block designs with easy-to-understand visual interfaces. These will help you get the best results from your business-analyst/developer teams as they work to discern what’s most valuable in the coming reams of IoT data.

Check out Vitria WHAT IF white paper to learn how service operations performance can benefit from an advanced analytic solution.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>