Last updated on November 10th, 2015 at 05:17 pm
There’s a lot of talk lately about predictive analytics, and undoubtedly it is starting to infiltrate the B2B software industry.
In the last year plus Salesforce.com acquired RelateIQ and LinkedIn bought out Fliptop, making meaningful progress in this space. But how will they implement predictive analytics solutions within their larger B2B cloud ecosystems? What approaches will even bigger enterprise software players—Microsoft, Oracle, SAP—and others take? Make, buy or partner? Industry analysts cannot answer definitively yet.
“Predictive analytics firms largely position themselves as platform and infrastructure rather than standalone applications,” says Kerry Cunningham, research director, demand creation strategies, SiriusDecisions, global B2B research and advisory firm. “There may be standalone applications, but the future probably lies in embedding predictive within other applications, such as CRM, marketing automation, web content management and programmatic ad buying.”
Not to say that CRM, marketing automation or other app makers will buy or build predictive applications, according to Cunningham. However, predictive technologies and the data science within them are probably more complex than the functions within which they would embed, Cunningham says.
“Simply building your own or buying your own to embed may not be the answer,” he says. “Plugging those systems into a predictive backbone or adding a predictive engine as a layer in the stack may be another way to think about it.”
“Every software company will have to think about predictive analytics,” says Jamie Grenney, vice president, marketing, Infer, who helps B2B companies predict which prospects become customers. “If you don’t have it you’ll need it.”
The way Grenney sees it? Salesforce and LinkedIn have taken the lowest common denominator approach to predictive analytics. In their models, it’s the easy way to predictive with sales automation, basically boiling down to whom to call next.
“We have a different model than others,” he says. “We’re connecting marketing automation with predictive.” Infer tries to predict who is how likely to buy and when, if ever, filtering out bad leads and surfacing good leads, Grenney says.
Integrated vs. standalone predictive solutions
While every B2B company may have to think about integrated predictive analytics, not everybody will make the same decision in the end. Right now, many companies build standalone platforms but use cases are always tightly coupled with other products.
“Predictive solutions will eventually be less of a standalone platform and more of an integration with products from within,” says Jason Maynard, director, data and analytics, Zendesk, cloud provider of customer service software. “This will be accelerated by companies like Amazon AWS and Google Services building (machine learning predictive) components into their developer platforms.”
In addition, these companies have an advantage because it’s more likely that machine learning predictive applications will be developed and tuned for specific utilizations with open source libraries and techniques, commoditizing the whole space, according to Maynard. Other platforms pursuing this strategy include Microsoft Azure.
However, not everyone believes that predictive analytics will become a commodity. Leaders in standalone predictive analytics see it differently, including Tal Segalov, co-founder and CTO, Mintigo, provider of a predictive marketing platform for accelerated revenue generation.
“…predictive analytics are not a standalone tool that is commoditized to the point where you don’t need to know much to use it,” he says. “It is not a panacea you put in and just forget about. What we are seeing in the big data world is that there are a few vectors in action.”
He goes on to say that predictive analytics are specific to particular domains, need strong data beyond email addresses in reference to customer companies, will provide more value to business leaders—sales, marketing, operations—than data scientists and must have wide availability.
“Predictive analytics will not become a general-purpose solution—a solution seeking problems—but rather a set of multiple solutions geared (to) and answering specific problems,” Segalov says. “We will see it more and more in every aspect of what we do, from marketing and selling to manufacturing, education and healthcare.”
Customer experience, cloud software and legacy vendors
An emerging category where predictive analytics finds a foothold comes by way of customer experience (CX) software. With the availability of customer big data, predictive analytics has become an important component of many cloud applications, including CX management software, according to Mark Magee, vice president, product management, MaritzCX, provider of customer experience software and services.
Traditional tech powerhouses have also seen this trend. But not without raising suspicions of their motives from native SaaS, pure-play predictive providers.
“We are seeing legacy vendors (e.g., SAP and Oracle) attempt to cloud wash their products or force-fit them (e.g., SAS for IoT) into existing application environments,” says Simon Arkell, CEO, Predixion Software, a predictive analytics platform vendor. “This is not the first time we have seen this approach, and it creates solutions that are very complex and expensive.”
Magee has also witnessed this trend. “Google is backing startup Automatic Statistician, and IBM backed the Watson project that promises a computer that can read natural language and generate a hypothesis. At MaritzCX, we created Spotlight data mining for simple, automatic analysis of customer experience data.”
For CX data analysts, that means not manually linking data and exporting it to an external tool for statistical processing to create visualizations for decision makers, according to Magee.
Visualizing the future of predictive analytics
Speaking of visualization of predictive analytics, that’s where the future lies, according to visionaries in the space. They will be leveraged by big data fans and lean data advocates alike, some say.
Predictive analytic tools and features will be a basic component of software releases, with key automated algorithms and visual reporting as standard features, says Pamela Veraart, senior vice president/group head, analytic practice, C2G Consulting. “Added-value features, driven by the depth of big data available, will factor into the type of predictions, insights and reports available.”
But the deluge of visual data could be overwhelming. Which in turn might cause the dreaded paralysis by analysis. To avoid this, solutions that can condense multiple visual predictive indicators into an at-a-glance, single-source-of-truth view could become killer apps.
For example, while access to data remains a top need, according to Bluewolf’s recent The State of Salesforce Report, the breadth of data visualization capabilities can easily eclipse a seller’s ability to consume.
“Applied appropriately, predictive analytics can consolidate many dashboards into a singular concept; i.e., you have an ‘at risk’ customer,” says Adam Bataran, senior director of analytics for Bluewolf, a global consulting agency. “Dashboards are not necessarily a prerequisite for making a data-driven decision. Predictive will bypass dashboards and integrate directly to workflow.”
Read our earlier post this week on the rise of predictive analytics
Main photo via Flickr (Creative Commons)