Predictive Modeling

  • April 17, 2013
  • Rodin_TheThinkerAbout ten years ago, I wrote a paper that predicted that analytics and social media would converge in CRM.  I believed this for two reasons.  First, I believed social media was inevitable though I had no idea what form it would take.  Facebook was not on my radar and might not have been invented yet, MySpace was something for kids, and Twitter had definitely not been invented yet.  But Plaxo and LinkedIn offered tantalizing glimpses of what was possible in business with social networking and I believed something of that ilk would eventually drive CRM.

    My second insight was that those social media applications would, of necessity, churn up a lot of data that would be useless unless we pushed it through an analyzer.  But if we did the pushing, like making sausage, out the other end would come valuable stuff.  What amazed me then and continues to amaze me is the rate at which knowledge is doubling thanks to all the data being churned up by social media.

    In 1982 Buckminster Fuller, the futurist, architect and some would say crank, published “Critical Path” a book in which he estimated knowledge doubling like this: Take all human knowledge up to the beginning of the Christian Era and call it one unit.  It took 1500 years for that knowledge to double.  Knowledge doubled again by 1750, the beginning of the Industrial Revolution, and again by 1900.

    Today, knowledge is doubling at a rate of between one and two years and IBM predicts doubling time will be down to 11 hours by 2014.

    All this doubling has many practical impacts the most important from my perspective is that those who don’t learn how to manage Big Data and the big information it generates in our own businesses will be toast.  Another of those impacts is that our ability to generate new knowledge is so prodigious that it will become increasingly difficult to generate knowledge that is unique to any person or business.

    We generate so much data today that it is possible by induction and other processes to infer knowledge that others might have and might think is proprietary.  On the other hand, though, failing at analysis will leave much information hidden in a morass of data.

    This all points to the need and even urgency attached to developing strategies for dealing with Big Data.  Last week at SugarCon 2013 in New York, I gave a talk on the subscription economy, one of my favorite topics.  In it I quoted some research from Gartner — by 2015 35% of the Fortune 2000 would derive some of their revenue from subscriptions.  And this from Aberdeen — only about twenty percent of companies studied had subscription businesses that appeared viable in that they had high customer satisfaction and renewals and their new contract value (NCV) was a positive number.

    So while Gartner might be right in its prognostication, it leaves much unsaid because revenue is not profit — even a bad subscription company can generate revenue while it’s going out of existence because it loses money on every transaction.

    So, how does a company become part of that top quintile?  Simple.  They develop metrics that they derive from customer data about use, payments and sentiment and relentlessly pursue them tying to optimize customer experience and involvement.  Needless to say, metrics are made possible by analytics — both the reporting kind and the predictive data modeling kind.

    There’s a boom happening in the analytics business these days with companies like GoodData, Totango and Gainsight among many others and I think smart companies — subscription or otherwise — ought to pay attention.

    That was my simple message at SugarCon but the not so simple reality is that most companies are not on the bandwagon yet.  They still don’t know what to do about Big Data and they aren’t exhibiting the needed curiosity to figure out that it’s time to get on track with subscriptions and analytics.

    Another finding that comes to me from long practical experience is that while knowledge might be doubling very quickly, our ability to apply it lags and I wonder and expect that a metric of new knowledge vs. applied knowledge would show a widening gap between what we know and what we do with it.

    Very shortly I’ll publish an ebook of interviews with five CMO’s of some fast growing companies.  What’s interesting about each of them is how these marketers have embraced Big Data and come up with strategies and metrics that better enable them to understand their businesses.  They know that the knowledge doubling that we are all caught up in isn’t some abstract concept, it affects them and it represents one of the great opportunities of this new century.

    Published: 11 years ago


    This post is part of an occasional series on the AppExchange as Salesforce.com celebrates the seventh anniversary of its launch.  The series will focus on some of the most interesting AppExchange applications of the last year.

    The TAS Group is anything but an emerging company and its Salesforce Platform native sales solutions are powerful tools for enterprise sales teams trying to improve.  This company has focused on sales methods and has been affiliated with Salesforce for many years.  But more than simply providing a paper-based sales methodology, TAS leverages predictive modeling based on data contained in its knowledge base — and in Salesforce — to coach people through their sales processes and help to make better forecasts.

    TAS AppExchange products are found under the TAS Dealmaker brand and include, Smart Opportunity Manager, Smart Account Manager and Smart Playbook and Forecast.  TAS carries on a theme from other successful companies in this series by leveraging predictive modeling to greatly improve what has historically been a manual process.

    There’s nothing more manual than managing a pipeline full of sales opportunities and determining next steps or forecasting which deals will close.  For many years sales people have succeeded with manual systems in part because there have been so many opportunities that, usually, there were enough to make quota no matter what happened in a particular deal.  But in a tough economy with intense competition, sales teams have learned they can’t just “wing it” they need to be able to prove the value of their propositions and that often means adherence to processes geared to demonstrating that value.  This is where predictive modeling comes in.

    There are three major advantages to the TAS predictive modeling approach over paper methods and simple rules based systems.  First, TAS has captured data about hundreds of thousands of sales processes which it has used to develop its knowledge base.  Using its patented coaching engine TAS Dealmaker can access this knowledge base full of past sales cycles and make predictions about current processes.

    Second, by capturing relevant data about each sales process from Salesforce,  Dealmaker can build probability models that coach a sales person or manager in the specific instance.  This provides added confidence that a deal will advance to the next step in the process or close or to a determination that the deal it is not ready to move ahead.

    Finally, TAS Dealmaker solutions operate at a level beyond simple rules.  A rules based solution might recommend that a sales person perform a particular action or behavior sequentially regardless of what conditions are present.  That might sound like a good idea, but it’s a one-size-fits-all approach that does not take into account the variables presented by a particular deal or sales person.  But, using its predictive modeling engine to coach sales people, Dealmaker takes into account all of the variables of the particular sales process before making its recommendation.  By tailoring its recommendations, TAS helps its customers to improve sales results.

    The Dealmaker suite uses Salesforce data and as most people know, users can add data fields to their Salesforce instances to provide more fields that are diagnostic of their sales processes.  Dealmaker works with the TAS knowledge base when it is first deployed but quickly assembles data from a particular company’s people and processes through Salesforce CRM, refining its models as it goes along so that in a short time its rules have adjusted and its predictions are incredibly accurate.  The result is more accurate forecasts, better close rates and improved revenue flow.

    The TAS Group is based in Seattle with offices in Atlanta, Dublin and Reading UK.

     

    Published: 11 years ago


    This post is part of an occasional series on the AppExchange as Salesforce.com celebrates the seventh anniversary of its launch.  The series will focus on some of the most interesting AppExchange applications of the last year.

    TOA Technologies is an AppExchange partner focusing on mobile workforce management, which provides an important extension to the Salesforce CRM solution set.  In the early days of CRM the category went by a different name — field service automation or FSA — which better reflected the fact that until then, most of the work we now associate with deploying and directing field service personnel was manual.

    While the initial automation efforts delivered many benefits, including reduced costs and greater productivity, those benefits have been largely absorbed and companies today are seeking greater refinement in a new set of benefits.  That’s where TOA Technologies comes in.

    From automation to management

    TOA has taken field service the next logical step from automation to management, which requires greater use of information.  The information that TOA leverages comes from collecting copious amounts of time related data and then analyzing it to derive performance metrics.  These metrics enable TOA to render accurate predictions for a business focused on dispatching people and equipment to customer sites as well as internal business sites that may require services or deliverables.  The result is better customer experience from better knowing when to expect the service call and how long it will take.

    TOA has a two-step approach to analyzing and predicting mobile workforce activity.  For each company it has helped since its founding in 2003, TOA uses predictive analytics based individual profile patterns to optimize each unique workforce.

    First, TOA collects time-based measurements about each member of a mobile workforce and the way that each completes different types of work.  Then, TOA’s ETAdirect software uses this learned information to make predictions about when appointments will occur and how long they will take. Because these predictions are based on actual, real-time information, organizations can effectively communicate appointment information to customers, with the confidence the field workforce will deliver on the service promise.  Because the system learns based on information unique to the individual and organization, it takes five days, on average, to learn the patterns of a given workforce in a way that generates results. , on average, to learn the patten waesu yding next steps with John/Yuvalconsumption)of day Thursday. ability However, the solution still achieves incredible efficiencies and results on day one, because TOA can input previously collected industry data (from its expertise in the industry) or apply basic patterns or knowledge provided by the organization. Therefore, the technology can optimize schedules for a good starting point even before it collects data from the field.

    The customer’s time is money

    The focus on time and efficiency is more important now than ever before because in today’s more competitive markets, customer satisfaction, and not cost, has become a focal point.  Customers have choices in who they do business with and if a product fails to measure up a customer might go elsewhere.  But in today’s more competitive marketplace the service experience goes under the same microscope and if a service experience fails to measure up it can also be reason for customer attrition.  TOA’s secret sauce is its understanding that a focus on time efficient processes results in both cost containment for the vendor and greater reliability and a better customer experience.

    So, by capturing data that enables the system to accurately predict how long it takes to perform a repair or move a truck from one customer location to another at a particular time of day, a service vendor can more accurately predict a service window and detect any slippage in the schedule.  Predictive scheduling enables the vendor to promise a smaller service window for the customer, freeing up the customer’s time in the process.  This is one reason that TOA has been selected by several cable providers, telcos and pay TV providers including Dish Network, Cablevision, Cox Communications and Virgin Media.

    Salesforce connection

    TOA Technologies has more than ninety enterprise class customers and more than 66,000 users of its solutions.  At any time there are more than seventy thousand mobile assets actively tracked by the system.  ETAworkforce, TOA’s integration suite, which went live on the AppExchange in mid-2012, brings together Salesforce CRM and field service management powered by TOA.  ETAworkforce also enables customer service representatives (CSRs) in the call center to book appointments the way a traveler might book a flight, as well as create work orders — all within the Salesforce ecosystem.  This was impossible before ETAworkforce was available.

    Benefits

    ETAdirect predicts how the unpredictable events that occur daily in any field service delivery will impact the rest of the day, thus empowering the organization to see and manage problems before they occur.  According to Vice President Of Channels and Alliances, Jeff Wartgow, “TOA’s ETAdirect solution manages 75 million appointments per year.  Also, “Any industry that requires having a service person arrive on site, and on time,” is a candidate for the TOA system including furniture delivery, retail and home healthcare to name a few.

    Tracking time is a great way to improve customer satisfaction and save money. Because TOA tracks data rigorously, it’s easy to come up with its own performance averages a representative sample includes:

    • Reduction of service wait window to one hour
    • 30 percent improvement in on-time performance and 98 percent customer satisfaction rate
    • 70 percent reduction in “Where’s My Service/Delivery?” calls to the call center
    • 47 percent increase in the rate of jobs completed each day
    • 40 percent decrease in miles driven per appointment
    • 20 percent reduction in time to complete jobs
    • 20 percent reduction in unnecessary visits to a customer home
    • 75 percent reduction in overtime

    Conclusion

    As long as there are reasons to bring products and services to customers there will be a need to manage costs and expectations to deliver optimal customer service experiences.  TOA Technologies has shown an understanding of what’s critically important in servicing customers at their locations.  It’s understanding what happens in the field, down to the individual employee, and using that information to predict when things will happen and how long they will take. Managing time has enabled TOA to elevate the discussion from service automation to real service management.

    Published: 11 years ago