I’ve been through these technology adoption cycles before. I used to work on JBOSS; I worked in the search engine space when it was hot.
At first, developers (like me) love their shiny new technologies, want to master it and build their own applications on top of it. Now nobody would ever dream of building their own CRM system on top of JBOSS, they’d just buy. It’s the same story with Business Intelligence tools. You’d never try to build a custom BI tool on top of a traditional RDBMS these days. You’d buy Tableau, Microstrategy, Pentaho or any other number of established BI tools that were purpose built for reporting on RDBMSs.
And here we are with Hadoop. We’re at the stage in its technology adoption cycle where people are so hungry to get the benefits of Hadoop, that they’re searching high and low and paying out the nose for the few and the proud who can hand-build them a way to make use of Hadoop. The problem is, that it won’t be long before they find themselves up a creek when their talent get recruited away for a higher paycheck and they’re left with spaghetti code they can’t make heads or tails of. It’s a story we’ve heard time and again from our customers. We want to prevent more organizations from getting burned like that, and we always have. Its in our DNA.
Our core engineering team and I have worked on Hadoop since its humble beginnings when it spun out of an open source project I worked on, called Nutch. We spent several years implementing custom Hadoop solutions at major brands before Yahoo! even got into the game. We did it so many times, that one day, we looked at each other and said let’s do this one more time, but build it at as product. On a Friday in November 2009, I fired my engineers. That Monday, I rehired them and we started building Datameer. That day, we had more than 40 years combined Hadoop experience, and the knowledge that we’d have this technology adoption cycle to work through. We set out knowing that what the market would ultimately want is a self-service big data analytics tool, so we started building.
The first thing we did was focus on making data integration — of any kind of data — easy. Today it’s a 3-step wizard with built in connections to more than 55 common structured and unstructured data sources.
Then, we focused on building a UI for analytics that anyone could understand, which led us to an excel-like spreadsheet interface with more than 240 pre-built point-and-click analytic functions.
A year ago, we did away with the traditional box-after-box style dashboards and re-imagined data visualization altogether. We gave our users a blank slate to freely design what we call Business Infographics. We put in all the tools you’d find in something like a Powerpoint and gave our users the freedom to design and annotate their reports or infographics as they saw fit, all while never severing the tie to the live data.
Six months ago, we introduced an app market, making it possible for anyone to bypass the building of custom analytics altogether.
Today, we are doubling down on our promise of making big data analytics on Hadoop self-service and a business user function with the introduction of Smart Analytics in Datameer 3.0. You can get the full details in our press release, or on our website, but in a single sentence, we’re giving subject matter experts like doctors, marketeers, or financial analysts a way to do actual data science with simple point and clicks. What once were complex algorithms are now buttons you can click that will “automagically” identify groups, relationships, patterns, and even build recommendations based on your data. A data scientist would call what we’re empowering business users to do ‘data mining’ or ‘machine learning,’ but we aren’t building a tool for data scientists. This is Smart Analytics.
Like I said. It’s the doctors who need to be able to analyze data to improve patient care, not the data scientists.