I was introduced to Alteryx, Inc. a few weeks ago and was very impressed by what I saw. It is an Analytics Application builder designed to enable everyone to have access to advanced analytics priced to be affordable and designed to be usable. Its great strength is its ability to integrate structured - that is the traditional IT transactional data - with less structured - that is information like location, test from call centres etc. - and bring them together to inform a decision. This week sees the next release, so what's new?
I am not going to claim that Alteryx is unique in its ability to handle unstructured data. I would, however, argue that the traditional analytics stacks are designed to handle transactional data and the unstructured data handling tends to be off to one side or grafted on, and is not integral to way the tools works. Equally, the new kids on the block that are making a lot of waves at the moment, the line of thought tools such as Tableau and QlikView, are great at visualisation, but are limited in the range of data sets that they can readily ingest in the decision making process. So, no tool is perfect and, as Gartner points out when including Alteryx in their latest BI quadrants, it is not a full BI suite, it is an Analytics Application builder. So we have a readily deployed Analytics Application builder that can take in data sets from a wide variety of sources and apply statistical analysis to them to help make better decisions.
In support of its design brief, the new release adds in some robust features that should consolidate its position as a leading provider of advanced analytics being used on a day-to-day basis by business people and not highly trained (and therefore by definition somewhat expensive) statisticians. Firstly there is support for cloud services to enable both the applications themselves and the results of analysis to be shared, in both public and private clouds, with full API support and SDKs being provided to enable that to be done in an open fashion. So now, when deploying to the cloud, things like single sign-in and identity authentication are addressed explicitly with robust functionality.
Then, in support of the notion that we have to add to the pool of data that is used to drive better decisions, they are now offering out of the box integration to a key range of data sources that are required to make the best decisions. These sources are from social media like Twitter, Yelp, Foursquare; Sales; cloud data like Salesforce, SharePoint and Amazon web services; and then the big data sources of Hadoop and the Hadoop persistent data store Hive, via an ODBC driver developed jointly with Map R. So all of that data can now be added to the transactional data from in-house sources to provide advanced analytics applications that can then, at the push of a button, be shared in the cloud.
A thing that I had not realised on my first look at Alteryx is that its statistical and predictive analytics engine is in now based on R. This is important, because the people who are at University today are not being taught analytics using SAS or SPSS as may have been the case 5 or 10 years ago. Now, everyone that I talk to has been taught using the Open Source R. So Alteryx is positioned to use the skills that are going to be most common, which is critical, because the more common a skill the more affordable it will be. So Alteryx is taking R as one of its routes to delivering analytics and it is then going to make that robust engine as easy to use as possible within its framework.
An example of how they are making things as simple as possible is that they are starting to roll out vertical industry-specific Analytics Application builders for key market segments, where the tools are preconfigured to provide the basis for most common types of analysis within a given vertical. So Alteryx will soon be offering a Analytics Application builder tailored to the needs of a telco to handle customer retention issues, integrating data from the BSS and OSS sides of the business to provide in-depth analysis of events that can lead to defection, and that is just one example. Other Analytics Application builders will then look at the needs of retail, with store location analysis as one of the typical capabilities.
Alteryx would appear to me to be a winning combination of capabilities and, hopefully later this year, it will start to become more widely available outside of North America. It is interesting that, in the Garner report, it is identified as being deployed to handle the most complex analytics on a regular basis in their survey. I think we could all do with access to that sort of analysis at an affordable price and without needing to employ an army of boffins to get there. I will continue to keep a watching brief on Alteryx, as one of the tools I think is most likely to break through as a big-time player. Those of us outside of the North Americas are probably already seeing Alteryx in applications coming from partners such as Experian so there is a solid basis for that growth in critical geographies like EMEA.