When someone at April’s AIIM16’s Industry Insights 2020 Expert Panel said, “We’ve been talking about text analytics for 20 years,” I wanted to jump up and yell, “So when are you vendors going to stop talking and do something?” At that point I decided it was time for me to take a deep look at text analytics.
The AIIM conference had a lot of vendors flagging “Text Analytics” but any review showed most were really just doing OCR indexing of forms and documents. Whoopee. I could do most of their demos using zonal OCR. Real analytics is not about finding a small piece of hard data. It’s about discovering an implausible insight buried inside one of the pieces of content.
The sample I’ve given before, built on Documentum’ case study:
EMC Documentum used to talk of its success with Pfizer’s Viagra. By using Documentum, Pfizer was able to submit their new drug application to the FDA 3 months earlier giving them 3 extra months before their new drug went generic. With a drug like Viagra, that meant millions.
It’s a great Enterprise Content Management (ECM) success story. It put a nice big number to cutting 3 months from document production. But there is a successful Text Analytics case hidden in that same scenario.
The real story of Viagra is that it was an accident. In clinical trials, the drug showed its side effect was more valuable than the original drug. In 2014 alone, Pfizer made over $150m on the sale of Viagra. How many other Viagra-like drugs are locked away in drug trials content collections that Text Analytics could find?
Text Analytics in ECM is missing a demo, a use case, an elevator pitch, heck a dream. We need to stop talking about developing yet another custom solution to file auto insurance claims. We need to have a vision for ECM that includes Text Analytics. Maybe it’s detecting fraud in those insurance claims.
I’m attending Text Analytics World next week. Here’s to hoping I find a good use case to promote to the ECM Community.