Stay marketing-savvy and tech-savvy. Get the latest in martech by subscribing to MarTech Today.
Super-Smart Business Intelligence Engine BeyondCore Partners With Microsoft To Offer a Free Version
The data analytics software is so smart, it decides which questions to ask and which results are most important.
For many marketers, the big obstacle in data analysis is that they’re not data scientists and don’t know the right questions to ask.
While data analytics software is getting friendlier to non-data scientists, San Mateo, California-based BeyondCore tries to take this even further by automating “the discovery of meaningful patterns in data.”
And today the company is partnering with Microsoft to offer a freemium version.
It is releasing BeyondCore Analyst for Office, which integrates into Word, Excel, PowerPoint or Outlook and allows ten analyses per month for free on datasets in Excel or Microsoft SQL. A premium service at $100/monthly allows unlimited analyses as well as the ability to query larger datasets, including big data ones in Hadoop, SAP HANA or Oracle.
The reason the company decided to release a freemium version, founder and CEO Arijit Sengupta told me, was because it kept hearing, “This is too good to be true.”
“People have gotten so used to vaporware,” he observed. He added that data analytics often rely on the user to find the most important results, while BeyondCore specializes in that task.
In its Office version, BeyondCore shows up as a side pane in a document from those applications, and it analyzes the data locally or via online.
A marketer or any business user first selects up to four variables from several dozen offered — such as revenue, store, discount or promo — and then chooses from a menu an initial focus for that variable. For instance, if the variable is “revenue,” you can choose an initial focus on what impacts revenue.
Then BeyondCore goes beyond.
Sengupta says the engine “looks at every possible question relating to that data.” It then decides which are the most important results to show you in graphs, tables and a brief narrative summary about why it chose what it did. Graphs can be descriptive, diagnostic or predictive, and suggestions are offered. New analyses can be generated from the initial results.
The user can modify the narrative and share the results document with others. Permissions can grant others the ability to run subsequent analyses, or limit them to view-only.
BeyondCore offers a case study of an anonymous retailer that it says used the engine to find “the levers, intrinsic behaviors and characteristics that impact customer value the most,” so as to find new and more efficient approaches to marketing campaigns.
It said the results showed the most lucrative micro-segments to target, high-value product combos and ways to increase coupon ROI. In all, it generated 16 analyses and 45 insights in one day.
For instance, the case study says, BeyondCore found that coupon use in general was resulting in a negative ROI, contrary to what the retailer thought. But it also found out that:
“ …new female customers who used the coupon, first purchased in the fashion specialty store, and made at least two purchases in the first two weeks had a net effect of $25—a return of 500%. A much more effective campaign could now focus on fostering this pattern to bring a much higher ROI.”
Similarly, it wanted to know why a marketing campaign on Facebook targeted at young customers in Florida was unprofitable. BeyondCore said its engine found that, again contrary to the retailer’s original analysis, young Florida customers were unprofitable but young customers in New York were “7.5 times more valuable than average.”
Sengupta says that the machine learning embodied in BeyondCore can be compared to IBM’s Watson, which famously beat human contestants on the TV show Jeopardy and is now making its living by providing expert advice across a wide range of industries.
BeyondCore’s software differs from Watson, he said, because it doesn’t need to know what you’re looking for, it does statistical testing on the quality of the data, and it offers versions that employ the already-familiar Microsoft Office interfaces as the front-end.
Founded in 2004, the BeyondCore team undertook eight years of research and development before releasing an initial version about a year ago.
In April, it launched its Apps for Microsoft Office 365 version. Sengupta said the new Analyst for Office differs because it has a free version, plus it can analyze big data sets from within Office documents.
Analyst is available online for Mac or PC, but the downloadable software version is available only for PC. Sengupta said there currently are no supporting APIs for Mac Office software.
Also offered is an Enterprise version, which includes an animated briefing of results and the ability for results to be narrated by a text-to-speech voiceover.