Analytics
News, events, thought leadership and more.
Mike Gualtieri, an analyst at Forrester Research, has just published a report on Big Data and Predictive Analytics: The Forrester Wave: Big Data Predictive Solutions, Q12013. If you are Forrester subscriber you can access the report here.
While it is good to continue to increase visibility on this topic, this report reinforces the core problem in Big Data analytics – specifically that Big Data analytics are the domain of the highly specialist Data Scientist, that unicorn-like role, that is most frequently attached to the topic of Big Data.
I should be very direct and say that Alteryx, while mentioned in the report (in a single sentence), is misrepresented as having a very limited role to play in Big Data analytics. We were told that the selection of vendors was based on previous years’ market landscape analysis for predictive. Alteryx customers like Southern States Cooperative, and Kaiser Permanente will be shocked to see that their market-leading efforts are not reflected in Forrester’s research, and instead the legacy, last generation of predictive tools have won out.
Ultimately, my problem with the report is not just that Alteryx does not receive sufficient credit for our Big Data access and analytical capabilities (which it should, along with other ‘modern’ analytics vendors in the market); instead it is that the report’s core approach fails to reflect the reality of the market or the needs of organizations. This reality is one where a single business analyst is asked to pull together data from multiple sources, including Big Data, and then within the same workflow deliver sophisticated analytics to the business – including predictive capabilities.
Here are some thoughts on why the report doesn’t help anyone except the legacy vendors below.
The Target User For Big Data Analytics Should Be The Business Or Data Analyst – Not Just The Data Scientist
The report’s Key Takeaways section actually mentions both user types (business analyst verses data scientist), but the vendors chosen for evaluation are almost entirely focused on the hardcore stats users. Despite including both user types, Forrester chose a range of vendors who rarely have tools that are ‘business analyst’ ready, while also highlighting this weakness in almost every vendor.
For example Forrester rightly points to the lack of business analyst-ready tools from SAS Institute, highlighting the need for it to for it offer tools that business analysts can actually use. Forrester even includes Oracle in this evaluation, despite their decision to not fully participate in the research as stated in the report, while criticizing their reliance on complex R scripting and then exhorting them to provide tools that are not complex. Salford Systems is also included despite not having a focus on business analysts.
One of the core Alteryx business benefits is to provide the line of business or data analyst with the data, analytics, and app production capabilities they need to meet the needs of their business – this includes the integration of Big Data and it includes accessible predictive analytics, all in a ...
Big Data Analytics Should Not Take 10 Tools And An Army Of IT Staff To Deliver What The Business Requires
This focus on specialist users also undermines the assessment of the technology and tools involved in the evaluation. When Alteryx is chosen by customers, it is often because tasks that took many tools, and multiple steps, are now achieved with a single tool, with much faster results. Even the table of products for each vendor included in the evaluation, undermines the idea of getting Big Data analytics to the masses. For example, to fulfill the data access, analytics and presentation of Big Data predictive analytics, SAS Institute has 26 tools listed, many on different product version numbers. This impacts the speed and cost of delivering analytics to the business. Alteryx customer Analytics IQ can be seen making that exact point here.
Having Predictive Capabilities Does Not Mean A Vendor Is Fit For Big Data Analytics
In the report, Alteryx is listed as an “embedded solution”. I would surmise this is because we have embedded R into our platform. Interestingly enough, vendors included in the report such as Oracle, Tibco Spotfire, and Revolution also have a reliance on R for their predictive capabilities.
The difference for us is that we focus on making the power of R more accessible to the mass of business analysts. We make R a first class citizen within our platform and have built out over 20 prepackaged R-based tools for the most common types of statistical and predictive tasks; ranging from customer clustering to time-series analysis.
Additionally, in the same workflow, we provide the capabilities to create the analytical dataset that the analyst needs. We don’t force them to ask the data warehousing group or some other specialist to go get it for them.
This means that someone can just drag and drop predictive capabilities into a workflow that also includes the access and integration of data from Hadoop, MongoDB, Salesforce.com, Teradata, or Microsoft Excel files. There is no need for scripting or multiple tools, and therefore, it is an extremely useful fit for Big Data analytics.
Conclusion
The publication of this report highlights the importance of the combination of Big Data and predictive capabilities in analytics, but does not help promote the market’s understanding of how and where businesses will get value from Big Data. Without the power of usable predictive, statistical and location intelligence, Big Data’s value becomes extremely limited.
I have a great deal of respect for Forester Research, and Mike Gualtieri in particular, but they have significantly missed the point of the growth of Big Data analytics and where its true value will be: in the hands of business decision makers supported by business analysts.
Paul Ross
VP of Product and Industry Marketing
Alteryx
