There's a huge gap between spreadsheets on the one hand and statistician- and data-scientist-oriented workbenches on the other. Alteryx is doing its best to fill the void.
"Excel needs a BFF!," quipped Alteryx president George Mathew at the company's May 18-21 Inspire15 event in Boston.
That line drew a big laugh from the more than 800 Inspire attendees because many Alteryx customers use its software to replace broken spreadsheet-based analyses. At Home Depot, for example, Charles Coleman, senior analyst, special projects, used to spend two weeks pulling together point-of-sale, marketing and merchandising data in Excel in order to study store clustering and performance. Presenting at Inspire15, Coleman explained how he now uses Alteryx to blend and analyze that data in less than an hour.
There are millions of analysts out there who are ready for something more powerful than Excel, yet most aren't ready for (or just aren't interested in) the coding and complexity associated with traditional data-mining and statistical-analysis tools.
Alteryx appeals to these would-be customers with a desktop designer tool that combines self-service data blending and coding-free predictive and spatial analysis. Desktops can be combined with the Alteryx server for companywide sharing of analytic applications and scheduled reports. There's also a cloud-based Alteryx Analytic Gallery where you can quickly provision browser and mobile-device access to apps, reports and visualizations.
Alteryx data-prep capabilities include everything from data extraction and cleansing to blending and enrichment. In many cases Alteryx is used exclusively for these functions in conjunction with analysis and data-visualization options like Tableau Software, and, more recently, QlikView and Qlik Sense. At Inspire, Alteryx president and COO George Mathew noted that the company has more than 300 joint customers with Tableau, including recent wins at Audi, EasyJet, EMC, and Johnson & Johnson.
I sat in on a session presented by Levi's exec Michelle Londeree, who detailed how she used Alteryx and Tableau to quickly create drillable dashboards. The dashboards replaced conventional PDF reports that were simply taking too long to maintain and modify using the company's conventional, IT-centric BI tools.
Many Qlik and Tableau users don't even realize that the people prepping their data are doing so with Alteryx behind the scenes. But with interest growing in forward-looking predictive analytics, both Qlik and Tableau are actively promoting Alteryx as a partner that can help their customers embed predictive and spatial analyses. I witnessed this partner love first-hand at the recent Qlik Connections conference in Dallas, where the vendor had both keynote mentions and how-to sessions on bringing Alteryx predictions into Qlik analyses.
Once the data blending is done, the Alteryx Designer workflow continues with an extensive menu of tools for predictive and spatial data analysis that you drag and drop into place and then configure without coding. Workflows conclude with report and visualization options, and you bundle everything up into an application that can be shared with business users through the Alteryx Server or cloud-based Gallery environment.
Alteryx' Next Release: In-Database and Spark
The next major release of Alteryx, due this fall, will bring upgrades for novices and experts alike. The newbies will get more tutorials and data samples at the desktop level to help them get started quickly. There will be charting upgrades and new analyses including social (graph) exploration. Server-level upgrades in the next release will include enhanced collaboration and version control, scheduling features, better auditing for data governance, and new scaling and redundancy features.
The next release will also bring enhanced data connections, with better REST ties to Salesforce, Marketo, MongoDB, SharePoint, and Redshift. New sources for Alteryx will include Qlik QVX files, JVX (JSON) reading and writing for Cassandra, and SAP Hana. In the big-data-analysis vein, Alteryx is developing in-database analytics ties with Teradata and Amazon Redshift and push-down analysis capabilities with Hadoop and Spark. These in-database and push-down approaches bring the analysis to the data rather than moving the data to the analysis -- a proven time and labor saver when dealing with data at scale.
Apache Spark is the darling of the conference circuit this year, frequently mentioned and often the subject of keynote appearances. Earlier this month at Informatica World, the guest was Professor Michael Franklin, director of the UC Berkeley AMPLab where Spark was invented. At Inspire15, Ion Stoica, CEO of Databricks, the commercial developer of Spark, joined George Mathew onstage. Both speakers touted Spark's in-memory performance and its broad array of analysis approaches, including SQL, R, machine learning, graph and streaming.
Spark integration will extend Alteryx support for R-language-based analyses to large data sets through the Spark R engine. Mathew also touted Spark as a better, faster alternative to MapReduce on top of Hadoop, and he promised Alteryx will become a "first-class citizen" within Databricks Cloud, that vendor's Spark-based big-data analysis environment running on Amazon Web Services.
MyPOV: Welcome to the Big Leagues
Alteryx is growing fast. Founded in 2010, Alteryx had only about 200 customer firms 18 months ago, but executives say that count will surpass 1,000 by this summer. Between its self-service data-blending and analytics capabilities and its Qlik and Tableau partnerships, the company is also winning bigger, broader deals with higher-profile customers.
With the coming in-database and big-data platform ties, Alteryx is going to draw even more attention, and, inevitably, more direct competition with some very large vendors. George Mathew made it sound like the gap between Excel and SAS/SPSS is a "greenfield" that Alteryx has all to itself, but giants including IBM, Microsoft, SAS, and SAP are playing close attention to this space.
The biggest competitor is surely SAS, which recently integrated its BI-oriented Visual Analytics and analytics-oriented Visual Statistics products. Both were introduced within the last two years to support coding-free, drag-and-drop analysis. SAP is converging its traditional data-mining workbench and its business-user-oriented KXEN InfiniteInsight acquisition in SAP Predictive Analytics.
IBM is aiming for a broad user base with Watson Analytics.
And Microsoft, too, is gearing up in this arena, building on predictive cloud services like Azure Machine Learning, acquiring Revolution Analytics, and getting more aggressive with its recent revamp of Power BI.
Make no mistake, self-service data-prep and self-service analytics are following in the footsteps of self-service business intelligence. So the gap between Excel and advanced analytics tools won't be Alteryx's alone to exploit. But Alteryx is doing a very good job of pioneering this space with an end-to-end, self-service workflow, plentiful data-access options including cloud sources, and a growing story around big-data analysis. Customers at the company's event seemed truly inspired by the capabilities that are there today and by what's coming.