The Next Generation of Data Analysis

site ai narrative

In the near future you will use voice commands to ask for actionable stories to be related to you from raw data. It sounds like Star Trek but it’s months away, not years.

Today is the start of IBM’s SmarterCommerce Global Summit 2013. As part of the program IBM invited several industry influencers to attend the event. Last night, I spent some time chatting with Sandra Zoratti, author of Precision Marketing, Stratigent‘s Bill Bruno and Triberr‘s Dino Dogan.

I shared with them how I was first exposed to real data driven marketing online in the early 1990s. I was managing banner ad placement ads on sites like Yahoo!, Excite and AltaVista for a consumer focused software company.  We could measure by each ad creative and placement how much usage and revenue they generated from their software download. From the start we developed hooks into the software to marry up the acquisition data. We could then ask the analysts to run queries for us and send us Excel spreadsheets. We used them to decide where to keep investing in banner ads.

So, of course, when Jeffrey and I started focusing more of our time on online businesses we expected people to do similar things.  Where we ever surprised!

Excel became the de-facto tool for analysts. Do you remember the days of running web logs files in Excel to analyze them? This was before tools like WebTrends log analyzer became popular and of course well before any of the javascript based web analytics tools even launched. Excel is still a great tool for analysts as Chief Evangelist for Bing Ads, John Gagnon has shown in his last couple of columns. It is not however, the best tool to share data  and collaborate with the entire organization. How many times have you seen Excel spreadsheets and charts go straight from the inbox to some digital black hole, never to be used again?

The only way to get people to use the data, is to make it more accessible to them. – Tweet this

This past week, we saw data visualization platform Tableau, enjoy a successful IPO. I love what you can do with their platform, but  it still doesn’t pack the punch it needs to in order to get an organization to act on the data. As Robbie Allen, the CEO of Automated Insights points out “most visualizations require the user go through the mental exercise of interpreting the results.”

This past February, I cautioned analysts that they have to up the value of their game to move up the value chain of their organization. At the time, I shared technology from NarrativeScience that transforms data into stories. They currently have a private data that I shared at eMetrics San Francisco that plugs in Google Analytics into their platform.

This morning, Automated Insights launched an impressive, similar service, Site AI. You can currently plug-in either Google Analytics or Clicky into their platform and receive daily emails that provide you with a cliff notes version of your analytics. The service is free to try and they are planning on pricing it based on how many people sign up. The more that sign up now, the lower the price is going to be. The technology is not perfect yet, but I see where it is going and I believe this is just what millions of website owners need today.

What really excites me about the Big Data scene is where the new interfaces for data are headed. Turning data into a story is awesome and having a computer discover insights and patterns is more efficient than having humans do it. However, if you look into the crystal ball, you may imagine the day that we have a Google Now voice interface into our data sources. Where you, I or your mom could go in and ask the data any question, to look for any pattern and make accessing data as simple as doing a voice search on your iPhone or Android device today. We’ll have predictive cards that inform us to look at new data points that should be of interest to us to further explore.

How far away do you think this might be? How will this lead to smarter commerce, smarter cities, smarter consumers? How else do you think we can humanize the data interface?

Please share if you think others would benefit.