Alteryx announced its new release earlier this year. Directions Magazine interviewed Paul Ross, vice president for product and industry marketing, to discuss how the more user-centric design of the 9.0 version allows for decision making closer to line-of-business managers, accomplishes fuller data blending, and allows direct integration of social media data feeds.
Directions Magazine (DM): In Alteryx's 2014 predictions webinar last December, panelists discussed how analytics would increasingly move into the realm of line-of-business (LOB) managers; data scientists might be less essential, as analytics tools become more "accessible" in terms of answering questions. How does Alteryx 9.0 support this with either additional functionality or improvements to the user interface?
Paul Ross(PR): At Alteryx, the mission of making analysts more productive by giving them deeper business insights faster has always been core to what we do, so Alteryx Analytics 9.0 continues this mission. Specifically, we focused on identifying the sources of data from which the analyst audience could be getting value about their customers but are prevented from doing so by the complexity of accessing that data. That is why we’ve added a lot more new and emerging connectors to data sources like social media aggregators, Google Analytics, Marketo, Amazon Redshift and JSON sources. Our users want and need this data to be blended with the rest of the data they have, without having to wait for a long data warehouse or Big Data project.
DM: We increasingly hear the term “data blending,” and in most cases it's a reference to merging structured and unstructured data or bringing together data from sources which might not normally be easily integrated. How specifically is Alteryx 9.0 a step forward in data blending?
PR: For our analyst users, data blending is about being able to create the analytic dataset they need to answer their question and to do it now – no matter what source it is coming from. Its not about just exotic datasets; for example Ingersol Rand/Trane uses Alteryx to bring together multiple survey datasets without relying on horrifically linked Excel documents or slow IT-led ETL projects. We already provide access to the widest range of datasets and types that analysts require and with 9.0, we are making it easier to access these emerging sources of customer insight.
A great example of the importance of data blending to organizational success is how Kaiser used Alteryx to shift away from a long manual process to blend data from across multiple systems that captured its customer (patient) experience of integrating with the organization http://www.slideshare.net/Alteryx/faster-efficient-data-blending-kaiser-alteryx. The problem this solves for the analyst is that they no longer have to manually do V-lookups in Excel or ask someone else to code data by location. This enabled them to not only unify the view of the customer but also to bring together data from across their geographically separated organization, providing corporate clients with data about which of their employees is covered at a national level.
DM: You mentioned the DataSift connector and GNIP in the launch press release. How do they integrate with Alteryx 9.0? How is social media integrated into the analytical workflow and then incorporated into location analysis?
PR: Until now social media data has been the purview of either hard-core data science projects or activity monitoring tools. By integrating the social media aggregation and processing capabilities of DataSift and GNIP we make this just another data source for an analyst. If a marketing department wants to be able to understand the return on their social media marketing investment in a specific location and compare that against the ad word spend, they can now do it in Alteryx. Each part of the workflow of getting from raw data to the business insight is supported: from blending social media data with customer data, to analyzing patterns at the ZIP Code or sales territory level, to delivering it in a format that the decision maker needs, including reports, analytics apps, Tableau or Qlik visualizations, or even writing back to a database.
For social media data, we have seen a big increase in the demand to access that data in customer, sales and marketing analytics. The ability to understand customer or prospect social media activity, down to a specific location, is becoming critical to understanding multi-channel marketing and sales. As a vendor, we had a choice – to create custom connections to every social media service or to create connectors to the aggregators who can provide us with even greater metadata, plus access to all the historical content (e.g. every Tweet ever). We chose the second option by integrating with DataSift and GNIP to make their data feeds yet another source that can be queried from within the Alteryx workflow. We will maintain some direct connections – for example, for Twitter – but those are more for live querying of the sources, not the historical content.
In terms of how this appears in our product, we have a customer connector that presents what is effectively a visual query interface - a simple-to-use approach that is very much in line with the rest of our spatial, data blending and predictive capabilities.
DM: With regard to “existing SalesForce.com integration Alteryx empowers analysts to blend data from across the full sales and marketing lifecycle” … and you’ve added similar connectors to Google Analytics and Marketo? Can you elaborate on how this supports analytics across the product lifecycle?
PR: The majority of our customers are using Alteryx to better understand their customers and to help make decisions that will impact them. This means a lot of data from marketing programs both inbound and outbound have to be taken into account. When we look at the analytics around these systems they are incredibly difficult to use or very siloed. Have you ever looked at the data structure for salesforce.com or the reporting in Marketo? It’s terrible for any marketing department that actually cares about whether its activity delivers revenue. We saw this gap and decided that 9.0 should really focus on the end-to-end process of getting customers’ attention (Google Analytics), to driving action (Marketo) to purchasing (Salesforce.com).
DM: Amazon Redshift, Pivotal, Greenplum and HP Vertica – these are databases, data warehouse and analytics platforms. Is Alteryx 9.0 an SaaS solution that can ingest data from each of these?
PR: Analysts in line-of-business groups like sales, marketing or operations rely on the data infrastructure that their organization has built up; it’s our job to make all of that easily accessible to them. Amazon Redshigt, Pivotal, Greenplum and Vertica are becoming a bigger part of that infrastructure for many organizations so we have made sure we optimize our support for each of them, so that they are just another source of insight for analysts. It doesn’t matter whether this data is in the cloud or on premise; Alteryx can access and blend this data with whatever else is necessary to build the analytical dataset the analyst needs.
DM: Alteryx 9.0 “unchains data from legacy systems” such as SAS and SPSS. How is this accomplished? How is R increasingly replacing legacy tools?
PR: Alteryx 9.0 allows our customers to add critical sources of data to their blending workflows as well as helping them unchain data from legacy SAS Analytics files. Organizations have spent millions of dollars creating SAS files that replicate data all over the place, while making it inaccessible once the laborious work has been done in SAS, so we decided to help them blend that with all the rest of their datasets. We can also write to SAS Analytics so if customers want to have a much faster, more analyst-friendly and much cheaper way to prepare data for analysis in SAS, we can help. We’ve also enabled these capabilities for SPSS.
The R predictive language is really gaining traction with statisticians and data scientists. It’s open source (and therefor a lot cheaper than SAS!), and more people are completing education courses that give them R skills. For us, R is the underlying predictive language, but for analysts it’s not about scripting. We have created over 30 predictive tools that are based on R but analysts just drag and drop into their workflow, no programming required. This puts predictive in the hands of more analysts without the long language learning process. As Alteryx customer Greg Bucko of Southern States Cooperative says about analytics: "Most people don’t need to be writing code here in the 21st century, especially people who are integrated into the business."