Let me make an educated guess: If you’re reading this post, chances are you’re getting ready to implement text analytics into your business. I’ll attempt a further guess: Your end goal is to make a certain workflow more efficient - maybe to save time or resources – or you’re working towards improved decision making – perhaps through customer analytics, open-source intelligence monitoring of your competition or supply chain, or through other forms of analytics. So let me share some considerations that could be useful as you embark on this journey.
Text Analytics is not something that happens in a box on a shelf somewhere. It’s a living system that requires your stewardship. It involves regular processes and activities that will require attention and know-how: a practice. The core of practice is a combination of expertise, common business sense, and change management. And there’s also a fourth aspect that I’m keeping for the end of this post.
The expertise that’s needed really depends on your particular goals but it will possibly include several of the following. A first area of expertise helps ensure your text analytics platform is aligned with your organization’s worldview. By worldview I mean the way your organization sees its business or activity: the topics and objects you pay attention to as a business, and their relationships. This is typically described in a thesaurus, taxonomy or ontology that translates it into an organized scheme that your text analytics platform can refer to. Two particular know-hows are required here. The first is knowing the business itself. This is the role of a subject matter expert that you can appoint based on their structured understanding of your business. The second know-how is in describing this worldview in a way that is understandable by the machine. That’s what a taxonomist (or ontologist) does. If you are starting up your text analytics practice, you may want to hire one, or at least, acquire some know-how in this area. Keep in mind, your business evolves over time and you’ll want your thesaurus or taxonomy to stay up-to-date as it does, so you’ll need those taxonomy skills on a recurring basis. A related expertise is computational linguistics which involves programming or training your platform to deploy your worldview. This expertise can be supplied by your platform vendor, by a third-party, or in some cases by your taxonomist. Lastly, deploying your platform into your existing information system may require some integration work as well. This is typically handled by your IT or their service providers.
Common business sense comes into play when deciding how to apply the technology to your business. Where can it add value and under what business case? As I mentioned above, some of the most common applications revolve around workflow efficiencies (i.e. increasing a certain business throughput) or improved decision making. But what’s the low-hanging fruit? Where can the technology be applied with the least effort and for the most impact? Applying common business sense will lead you to the answers that make the most sense in your own business. On the other hand, if you don’t apply business sense (and I’ve seen it happen), your practice will falter.
The third aspect of practice is really change management. The reason is, in the best of cases you’re doing something entirely new, and in the worst case you’re changing the way things have been done up to now. People need to get accustomed to the new thing, and they need to participate. Pilot projects can double as internal showcases that help spread an awareness of what the new possibilities mean for the organization.
If you’re still with me, I see a fourth essential aspect in Practice – maybe you have it in mind as well? It’s workflow. Because that’s what happens when you’re successful: After you’ve assembled the team, chosen your business cases, translated your worldview into a taxonomy, managed change, all of it becomes a regular workflow, like the rest of your operations. And what you want is a smooth workflow that actually produces the results you had in mind when you were planning your practice, and that produces them efficiently. So don’t hesitate to ask your vendors: How easy is your platform? How fast can users get up to speed? What steps did you take to make their work simpler? I’m pretty sure you’ll quickly see the difference between vendors for whom your success is a daily obsession and those for whom it’s an afterthought.
In closing, I’d like to congratulate you for starting your Text Analytics practice. It’s a transformative technology and applying it to your business is a smart move. Needless to say, we’re still discovering new applications every day. What I find most exciting isn’t what has been done up to now: it’s what you’ll do with it tomorrow. And I’m really curious to learn where you want to go with it.