Business intelligence is firmly back on the public agenda for 2014 with nods towards the so-called 'big data' topic. We're seeing this through the steady stream of briefing requests, events and announcements coming out from the vendor community along with close scrutiny of earnings announcements. This is supported by field feedback that confirms a significant number of businesses are ready to move beyond technology refresh to intelligence based decision making.
Since 2010-11 it has became self evident that the business intelligence market was shifting radically. Self-service in the hands of users has been pushed by QlikTech and Tableau with direct impact on incumbent players like the SAP owned Business Objects, Oracle owned Hyperion and IBM owned Cognos/SPSS. It was very clear we had a disruption hitting all adjacencies in the business intelligence stack.
The entire data management field has started to be disrupted via interest and take up of Hadoop in its various forms - CloudEra springs to mind.
And let's not forget those who come from the so-called noSQL part of the market - MongoDB and Hortonworks being the vendors that are top of mind. Advanced analytics such as SPSS and SAS which have dominated the market for some year are the next targets.
When I was in the US, I took briefings from both Numerify and Alteryx that cover different aspects of this broad topic. If these are new names to you then that should not surprise. The last few years has seen a number of relatively new names coming to prominence.
NumerifyNumerify has correctly spotted that the move to all things cloud opens up an opportunity to be the Swiss Army knife of data aggregation. It is starting with IT, using the strapline: 'measure like a CFO so you can manage like a CEO' for the business of IT and concentrating on the fact that data will come from both cloud and on-premises systems.
The solution leverages ServiceNow which focuses on managing IT assets and landscapes. To give some idea what this means, the company described three examples of problem solving inside customers' landscapes:
Customer 1 - They want to understand the true cost of servicing an incident. How often do incidents gets passed from one agent to another - what are they? Who is solving the most problems. The ideas is to uncover patterns where the flow of resolution has choke points.
Customer 2 - Wants to know the mean time to resolution in 'business time' across different geographies. This then becomes the starting point for discovering where best practices are arising and in turn, how to push those across all operating environments.
Customer 3 - What’s the most costly incident type to service? Data has to be collected from multiple sources and it’s collection has to be process driven. This is the best way to know the type of outcome and can get the right data types and data composition to arrive at the most informed results.
Why does any of this matter. If you are relying on a CRM system like Salesforce, it is easy to spin up extra columns to add in data you want to capture. But that means the BI side is almost always broken. Numerify takes that pain away by autodetecting what's happened and adjusting the BI model accordingly.
According to Gaurev Rewari, CEO, Numerify customers are getting better insights: "Some answers come right away - one customer saw peaks and valleys of incidents but they were contrary to what was expected. In this case there was an incorrect staffing problem. In a different case - a self-service customer, they found that despite having self-service, the walk up service was still prevalent - especially in the upper echelons of the company. That was contributing to costs they previously could not understand."
AlteryxAlteryx drew a lot of attention late last year when it landed a $12 million round. What's it been doing with the funds?
Alteryx has largely felt that the so-called data science capability is in the hands of a few number of people. These are largely PhD stats people who are data munging in a programmatic manner.
They then bring that aggregated data into an analytic model that is hand coded or uses something like SAS/Clementine. The end user story is then told around Excel. There are only about 200,000 data scientists who could do that. Hence it is no surprise that the frenzy around hiring data scientists is coming up short.
Alteryx believes that a big part of the problem centers on extract, transform and load (ETL.) They're right. I well recall seeing the Cognos ETL layer back in the 1990's and watching as it sucked all the joy out of understanding customer data, regardless of the fact that the final outcome was both beautiful (then) and useful.
To that end, Alteryx is building the ETL platform it believes is needed for the 21st century where self-service and end user prepared modeling is the topic du jour. That means aiming squarely at SAS as a competitor and, like Numerify, tackling the problem of blending multiple data sources.
The company believes it will build upon past 30% year over year growth to reach 40% growth this year but it is not thinking that will lead to large scale SAS replacements any time soon.
Its latest release, Version 9, focuses upon trade area analytics. McDonalds uses Alteryx to to open up its entire China growth strategy. They, along with many customers are trying to solve problems in a highly iterative way. According to George Mathew. president and COO Alteryx: "You have to work with software that responds to the the speed of users."
How did Alteryx manage to start encroaching upon SAS's turf? The company reverse engineered the SAS (and SPSS/IBM) file formats to get at the data. This was done in a clean room so there can be no question of breaking any copyrights or patents.
It's mantra? 'data scientists not required.' I'm sure the other marketers and buzzword lovers will hate that but it represents the ind of reality check the market needs.
These are early days in the next wave of analytics. Both companies believe that while the market makers, marketers and anal-ysts might like to talk 'big data' the real action is not in finding data analysts but supporting business analysts. I find this approach far more appealing than some of the outrageous claims that sound far too much like 'solving world peace' than they do about focus on business outcomes.
The practical approach of recognizing many different data types is welcome. As we have said in the past, just because stuff is going cloud doesn't remove the need for integration. If anything it accentuates the problem.
It is good to see both companies operating an 'expand and land' approach rather than try convince the world they are magicians. We will be following up with an analysis of how Unilever is achieving results from Alteryx.