Enterprises have lots of solid choices for big data predictive analytics.

That’s the key takeaway from Forrester's just released Wave for Big Data Predictive Analytics Solutions for the second quarter of 2015.

That being said, the products Forrester analysts Mike Gualtieri and Rowan Curran evaluated are quite different.

Data scientists are more likely to appreciated some, while business analysts will like others. Some were built for the cloud, others weren’t.

They all can be used to prepare data sets, develop models using both statistical and machine learning algorithms, deploy and manage predictive analytics lifecycles, and tools for data scientists, business analysts and application developers.

General Purpose

It’s important to note that there are plenty of strong predictive analytics solution providers that weren’t included in this Wave, and it’s not because their offerings aren’t any good.

Instead Forrester focused specifically on “general purpose” solutions rather than those geared toward more specific purposes like customer analytics, cross-selling, smarter logistics, e-commerce and so on. BloomReach, Qubit, Certona, Apigee and FusionOps, among others, are examples of vendors in the aforementioned categories.

The authors also noted that open source software community is driving predictive analytics into the mainstream. Developers have an abundant selection of API’s within reach that they can leverage via popular programming languages like Java, Python and Scala to prepare data and predictive models.

Not only that but, according to report, many BI platforms also offer “some predictive analytics capabilities." Information Builders, MicroStrategy and Tibco, for example, integrate with R easily.

The “open source nature” of BI solutions like Birt, OpenText and Tibco Jaspersoft make R integration simpler.

Fractal Analytics, Opera Solutions, Teradata’s Think Big and Beyond The Art and the like also provide worthwhile solutions and were singled out as alternatives to buying software. The authors also noted that larger consulting companies like Accenture, Deloitte, Infosys and Virtuasa all have predictive analytics and/or big data practices.

In total, Forrester looked at 13 vendors: Alpine Data Labs, Alteryx, Agnoss, Dell, FICO, IBM, KNIME, Microsoft, Oracle, Predixion Software, RapidMiner, SAP and SAS.

Forrester’s selection criteria in the most general sense rates solution providers according to their Current Offering (components include: architecture, security, data, analysis, model management, usability and tooling, business applications) and Strategy (components include acquisition and pricing, ability to execute, implementation support, solution road map, and go-to-market growth rate.) Each main category carries 50 percent weight.

Leading the Wave

IBM, SAS and SAP — three tried and trusted providers — lead this Forrester Wave:.

IBM achieved perfect scores in the seven of the twelve criteria: Data, Usability and Tooling, Model Management, Ability to Execute, Implementation Support, Solution Road Map and Go-to Market Growth Rate. “With customers deriving insights from data sets with scores of thousands of features, IBM’s predictive analytics has the power to take on truly big data and emerge with critical insights,” note the report’s authors. Where does IBM fall short? Mostly in the Acquisition and Pricing category.

SAS is the granddaddy of predictive analytics and, like IBM, it achieved a perfect score many times over. It’s interesting to note that it scored highest among all vendors in Analysis. It was weighed down, however, by its strategy in areas like Go-to-Market Growth Rate and Acquisition and Pricing. This may not be as a big problem by next year, at least if Gartner was right in its most recent MQ on BI and Analytics Platforms Leaders, where it noted that SAS was aware of the drawback and was addressing the issue.

SAP’s relentless investment in analytics pays off,” Forrester notes in its report. And as we’ve reiterated many times, the vendor’s predictive offerings include some snazzy differentiating features like analytics tools that you don’t have to be a data scientist to use, a visual tool that lets users analyze several databases at once, and for SAP Hana customers SAP’s Predictive Analytics Library (PAL) to analyze big data.

The Strong Performers

Not only does RapidMiner’s predictive analytics platform include more than 1,500 methods across all stages of the predictive analytics life cycle, but with a single click they can also be integrated into the cloud. There’s also a nifty “wisdom of the crowds” feature that Forrester singles out; it helps users sidestep mistakes made, by others, in the past and get to insights quicker. What’s the downside? Implementation support and security.

Alteryx takes the pains out of data prep, which is often the hardest and most miserable part of a data worker’s job. They also offer a tool for that helps data scientists collaborate with business users via a visual tool. Add to that an analytical apps gallery to help users share their data prep and modeling workflows with other users, and you’ve set a company up with what it needs to bring forth actionable insights. While Alteryx shines in areas like Data, Ability to Execute, and Go-to-Market Growth Rate, there’s room for improvement in Architecture and Security.

Learning Opportunities

Oracle rates as a strong performer, even though it doesn’t offer a single purpose solution. Instead its Oracle SQL Developer tool includes a visual interface to allow data analysts to create analytical workflows and models, this according to Forrester. Not only that, but Oracle also takes advantage of open-source R for analysis, and has revised a number of its algorithms to take advantage of Oracle’s database architecture and Hadoop.

FICO, yes, Forrester’s talking about the credit scoring people, has taken its years of experience in actionable predictive analytics, built a solution and taken it to the Cloud where its use is frictionless and available to others. It could be a gem for data scientists who are continuously building and deploying models. FICO’s market offering has lots of room for improvement in areas like Data and Business Applications, though.

Agnoss aims to make it easier for non-data scientists to get busy with predictive analytics tools via support services and intuitive interfaces for developing predictive models. While the solution provider has focused its go-to-market offerings on decision trees until recently, it now also offers Strategy Tree capability, which helps advanced users to create complex cohorts from trees.

Alpine Data Labs offers “the most comprehensive collaboration tools of all the vendors in the Forrester Wave, and still manages to make the interface simple and familiar to users of any mainstream social media site,” wrote Gualtieri and Curran in the report. The fact that not enough people buy Alpine products seems to be the problem. It might be a matter of acquisition and pricing options, it’s here that Alpine scores lowest among all vendors in the Wave.

Dell plans to go big in the big data and predictive analytics game. It bought its way into the market when it acquired Statistica. “Statistica has a comprehensive library of analysis algorithms and modeling tools and a significant installed base,” say the authors. Dell scores second lowest among Wave vendors in architecture, so it has a lot of room for improvement there.

KNIME is the open source contender in Forrester’s Wave. And though “free” isn’t the selling point of open source, it rates; perhaps only second to the passion of its developers. “KNIME’s flexible platform is supported by a community of thousands of developers who drive the continued evolution of the platform by contributing extensions essential to the marketplace: such as prebuilt industry APIs, geospatial mapping, and decision tree ensembles,” write the researchers. KNIME competes with only Microsoft for a low score on business applications and is in last place, by itself, when it comes to architecture. It has a perfect score when it comes to data.

Make Room for Contenders

Both Microsoft and Predixion Software bring something to the market that others do not.

They seem to be buds waiting to blossom. Microsoft, for its part, has its new Azure Machine Learning offering as well as the assets of Revolution Analytics, which it recently acquired. Not only that, but the company’s market reach and deep pockets cannot be overstated. While Microsoft brought home lower scores than many of the vendors evaluated in this Wave, it’s somewhat forgivable because its big data, predictive analytics solutions may be the youngest.

Predixion Software, according to Forrester, offers a unique tool, namely (MSLM), a machine learning semantic model that packages up transformations, analysis, and scoring of data that can be deployed in.NET or Java OSGI containers. “This means that users can embed entire predictive workflows in applications,” says the report.

Plenty of Good Choices

The key takeaways from Forrester’s research indicate that more classes of users can now have access to “modern predictive power” and that predictive analytics now allow organizations to embed intelligence and insight.

The analysts, of course, suggest that you download their report, which, in fact, might be worthwhile doing. This is a rapidly evolving market and vendors are upgrading their products at a rapid clip. We know this because there’s rarely a week where a new product announcement or feature does not cross our desks.

And if it’s true that the organizations who best leverage data will win the future, then working with the right tools might be an important differentiator.

Creative Commons Creative Commons Attribution-No Derivative Works 2.0 Generic License Title image  by Andrew E. Larsen.