Artificial Intelligence and Machine Learning Threatens to Knock Salesforce from its Perch, but it’s What Happens Next That’s Truly Astonishing
It was a balmy Monday in July, 1916, and thousands of nattily dressed professionals had lashed through the Beaux Arts front doors of the Arcadia Ballroom in Detroit to hear Woodrow Wilson give a fiery address to the nation’s sales force. More than 5,000 attended the president’s keynote during the weeklong World’s Salesmanship Congress, including Henry Ford and his sales director Norval Hawkins, as well as the phrenologist Grant Nablo, who, through dubious cranial research, recommended hiring salesmen with “high foreheads,” denoting acumen.
But among the young professionals of the National Cash Register Company and Burroughs Adding Machine, who attended seminars with patriotic names like “Uncle Sam’s Salesman,” a lesser-known innovator, still years from making his mark on the sales industry, was quietly roaming the conference halls with a cutting-edge agenda.
During a lightly attended seminar, publisher John Cameron Aspley, who had only recently begun hawking a series of mimeographed sales bulletins, implored savvy managers to harness the power of statistics to hit sales quotas and raise revenues. A year later, Aspley would go on to launch the wildly successful Dartnell Corporation.
“[He] instructed managers to encourage salesmen to gather information on their rounds and prompted them to do so with forms, questionnaires, and other tools,” wrote historian Walter A. Friedman in the 2004 book Birth of a Salesman. “Dartnell even suggested companies arm salesmen with cameras to take photographs of both the interior and exterior of customers’ stores. The photos could be studied to see what inventory the store carried and what room was available for promotional display, and, eventually, could be framed and given back to the customer as a gift.”
In the century since Aspley’s pivotal innovations and subsequent research, the so-called “science of sales” has evolved at breakneck speed. Ledgers chock-full of analytics that the Dartnell Corporation once evangelized made way for 1950s-era Rolodexes, dog-eared and packed with information on sales prospects. The advance of personal computing in the 1980s slicked the rails for a rise in software development that inspired market contact management solutions from companies like Act! and GoldMine, to name a few early disruptors. Such breakthroughs led to the advent of customer relationship management (CRM) software and the rise of sales force automation, thanks to pioneers like Brock Control Systems and Siebel Systems, whose founder, Tom Siebel, fled Oracle in 1990 after failing to convince cofounder Larry Ellison to market its own internal sales application.
“Sales professionals detest CRM,” said Messer. “For years it was perceived as laziness. After decades of its use as the primary tool to manage sales teams, it’s clear they saw the bigger picture—CRM doesn’t impact outcomes, it simply records them through a deeply flawed process.”
But the seismic technological shift that culminated in 1999 with the launch of Salesforce.com, among the first widely used Software-as-a-Service vendors, has largely flatlined, due in part to faulty predictive analytics and tedious data entry that, by one estimate, accounts for a whopping 20 percent of a salesperson’s workweek. In fact, analysts believe the accuracy of such technology is so questionable that it necessitates entirely new processes and people to extract any meaningful insights.
“Almost four generations of people have invested enormous amounts of time and money in trying to figure out how to make CRM work, but yet the average forecast accuracy rate is 46 percent, which is worse than flipping a quarter,” said Stephen Messer, the co-founder of tech firm Collective[i]. “It’s worse than being a gambler in Vegas, which is a 49 percent chance of winning. You have a 46 percent chance of being accurate on a weekly guess on whether you’re going to close deals next week. The problem is neither human nor technology. It’s the reliance on CRM as the primary platform for managing sales.”
Like phrenology and the questionable cranial research that Nablo espoused at the turn of the century, the once high tide of customer relationship management software is facing its ebb. According to a 2015 study by Forrester Research, 52 percent of Salesforce customers surveyed in a poll reported that the “high cost of ownership over time” ranked chiefly among a litany of problems associated with the San Francisco–based company’s Sales Cloud module. Meanwhile, 43 percent of those polled warned they would “renegotiate our contract with Salesforce when it comes up for renewal,” and another 8 percent said they would shift to a different service.
But rising costs, it seems, may only be the tip of the iceberg for Salesforce, with a current market cap of almost $70 billion. A reluctance to integrate its marketing and sales products—which employ separate user interfaces—has also prompted a sales disparity in its Marketing Cloud module.
But while high cost and functionality remain a bone of contention for Salesforce customers, the data entry debacle is endemic across a sprawling field of customer relationship management platforms, including those offered by Oracle, SAP and IBM. In a recent ranking of the best sales force automation solutions, Forrester Research described the process as an affliction infecting nearly every product on the market.
“The biggest pain point around legacy SFA has been its negative impact on productivity, thanks in large part to the amount of manual data entry required by end users,” according to the report, “Sales Force Automation Solutions Q2 2017.”
And the challenge isn’t lost on salespeople. “Sales professionals detest CRM,” said Messer. “For years it was perceived as laziness. After decades of its use as the primary tool to manage sales teams, it’s clear they saw the bigger picture—CRM doesn’t impact outcomes, it simply records them through a deeply flawed process.”
Despite the well-established elephant in the room, however, some of the Fortune 500’s most overzealous operations managers and executives have gone to absurd-seeming lengths to enforce the daily grind of data input—like withholding pay from otherwise capable sales professionals who repeatedly neglect the task.
“I want you to imagine, if you’re a salesperson, that you’ve been driving with a rearview mirror, and for your entire sales career, you’re used to a few bumps in the road, or you’re used to a few deals sideswiping you at the end of the quarter that don’t happen.”
Using sales professionals’ data input to track how a sale is proceeding is not only tedious and time-consuming but guaranteed to be inconsistent, biased and inaccurate, said Messer. The so-called garbage-in, garbage-out problem is pervasive and necessitates that sales managers conduct weekly in-depth inspections or pipeline reviews—in other words, face-to-face meetings designed to identify flaws and ensure the accuracy of faulty CRM. A separate team, sales operations, is devoted to cleansing CRM data before it can be analyzed for macro trends of sales activity.
“CRM is like a house of cards—every process, every decision, is based on how sales professionals accurately transcribe, or fail to transcribe, what happened in each of their interactions,” said Messer. “As a result, companies create an illusion of being data driven when the reality is that the majority of their sales organization’s time is spent collecting and correcting data, not acting on what it can tell you.”
The future, then, may hinge on applications that employ data collected from automated sources and threaded with artificial intelligence to allow sales professionals to spend more time on customer-centric selling, analysts predict.
Mary Shea, a sales analyst at Forrester, the market research firm, said the tipping point will be accelerated by the emergence of millennials in the workplace, not to mention a rising tide of Generation Z–aged interns at the periphery of professional life. Shea predicted the expectations of modern workers will hasten the demise of CRM, as employees raised on social media networks increasingly demand more innovation from their digital solutions. Indeed, by 2030, according to the Bureau of Labor Statistics, nearly 75 percent of the U.S. workforce will be represented by millennials reared from an early age on consumer-facing networks that deliver streaming insights rather than merely a database for sales management. Think Uber, Amazon, Facebook, Linkedin, Spotify and Waze, to name just a few.
“They’re going to want a much more seamless, user-friendly interface than, I think, CRM was ever designed to be,” Shea said of the coming onslaught of young sales employees. “In a vacuum, CRM was never intended to be the workhorse it is today, and I think, ultimately, users of the future are going to demand a digital interface that feels like LinkedIn or feels like Facebook, but for something that can manage all of their data input as it relates to operating or running their business.”
For the indomitable six-foot-five founder of Salesforce, such widespread flaws were never intended. When he was 15, Marc Benioff founded Liberty Software in his native San Francisco, churning out early video games, with names like Escape from Vulcan’s Isle for the Atari 800 computer. The start-up was successful enough that, in 1982, he was able to afford tuition at the University of Southern California.
It was following college, during a rapid rise at Oracle, however, that Benioff hit upon the idea of providing access to business applications from the web rather than software that companies were required to install. In the flash of an eye, the era of downloads and CD-ROMs had vanished. Cloud computing had entered the lexicon.
Since the turn of the 21st century, Salesforce has spun off nine product categories and invested heavily in social media, marketing and mobile technology. Its Salesforce 1, App Cloud and Work.com platforms have helped boost the company to $8.9 billion in annual revenues while swiping 45 percent of the sales-software market.
In the past five years, meanwhile, the company has been on a tear, acquiring the engineering team behind start-ups like MetaMind and spending a whopping $390 million for a suite of e-mail and calendar products developed by RelateIQ, a Palo Alto–based start-up with fewer than 100 employees. Perhaps most buzzed about, however, was the unveiling last year of Einstein, the company’s foray into artificial intelligence, which Forbes described in a 2016 profile as “a new nerve system across the entire business.” To say it was ballyhooed is an understatement, to say the least.
Many experts believe Salesforce and Benioff have lost their way. They argue the company has a lack of vision, buying their way into markets rather than creating them, citing failed efforts at IoT, the cloud and most recently, artificial intelligence. “Kudos to Salesforce for reaching maturity and bringing the cloud to the enterprise,” says Daniel Flax, a former chief technology officer of TheStreet.com and an award-winning CIO. “However, maturity has meant that they are no longer innovating in meaningful ways. It’s a classic innovator’s dilemma. Like so many other large organizations before them, they’ve moved into the role of innovation through acquisition which is also tricky. The value of CRM as a tool has reached a commodity stage with feature parity across most vendors and incremental value being provided around the edges.”
Early reviews of Einstein have been mixed, at best. Salesforce executives themselves have acknowledged that the effectiveness of its deep learning tools will rely heavily on the data it has at its disposal—so, in other words, the same data source managers and sales operations spend the bulk of their time repairing. Perhaps most concerning, however, is that for many in the sales industry and beyond, the market has shifted from cloud computing to intelligence, delivered by networks that offer state-of-the-art applications on top of massive data sets. Naturally, the promise of artificial intelligence can’t be fulfilled via antiquated software or isolated data sets.
“In other respects there isn’t a lot there,” Todd Berkowitz, an analyst with the research firm Gartner, said during an interview with the technology news site eWeek shortly after its debut in October 2016. “The predictive lead scoring isn’t utilizing a lot of external data sources you’d need to do it in the marketing automation platform, and it’s not nearly as sophisticated as stand-alone solutions.”
Rising discontent with Salesforce and software-hinged CRM, meanwhile, has failed to dull the rapidly growing enterprise intelligence market, which is expected to exceed $22 billion by 2020, according to Gartner research. Certainly, as the volume, velocity and variety of information collection grows at a staggering rate, the necessity to convert large data sets into valuable insights has relegated Salesforce and other CRM vendors to the backbench.
In anticipation of the sea change toward business intelligence, Microsoft last year repositioned its enterprise resource planning (ERP) and CRM platforms under a single umbrella, Azure, and disaggregated those components into myriad purpose-built applications. Buttressed by more than 2,600 third-party applications and the $26 billion acquisition of LinkedIn—a move many saw as devastating for Salesforce—the streamlined platform has drawn attention from analysts, who have praised Microsoft’s sophisticated pivot to enterprise intelligence. To be sure, in a poll of 235 information technology and application development professionals conducted by the machine data analytics service Sumo Logic, 66 percent named Azure as its vendor of choice—beating Amazon Web Services by a staggering 10 points.
“I’m incredibly excited about where this business is going,” said Alysa Taylor, a general manager with Microsoft’s business applications division. “I believe very deeply in this vision, and I think that Microsoft in particular has a set of assets across our cloud portfolio that no other vendor has. So bringing together our productivity tooling, our AI and data solutions, our infrastructure and apps into an end-user environment that surfaces through business applications is something that I don’t see any other vendor being able to provide.”
At the Flatiron District offices of Collective[i], meanwhile, Messer and his colleagues have quietly amassed what could be the world’s most robust B2B sales networks. Indeed, since launching in 2008 following the $425 million sale of the online distribution network LinkShare, Messer and his cofounders have strived to harness artificial intelligence and predictive analytics in service of a modern digital business network that, if successful, could thoroughly modernize the sales industry.
Messer equated enterprise intelligence and the predictive analytics capabilities of applications like Collective[i]’s network—aptly named Intelligence.com (see “The Intelligence Quotient,” right)—to driving a car. Unlike CRM, in which sales professionals manually enter the odds of an opportunity succeeding based on past activities, Intelligence.com connects directly to data sources like e-mail and calendar applications, altogether eliminating the need for data entry. Company executives believe these data sources, when aggregated and analyzed with technologies like natural language processing, can guide sales professionals to better outcomes and invaluable insights into buying behavior.
For Messer, the challenge to CRM is bigger than the flawed data source it relies on. The new paradigm is to observe buyer behavior via troves of data collected from tools routinely used by salespeople in their day-to-day activities that are aggregated and combined with machine learning to feed sales professionals real-time intelligence external to their organization, he said. Unlike CRM—a compilation of how internal sales teams recount the past—modern sales applications, like Collective[i]’s Intelligence.com, enable sales teams to improve future outcomes, Messer said.
“I want you to imagine, if you’re a salesperson, that you’ve been driving with a rearview mirror, and for your entire sales career, you’re used to a few bumps in the road, or you’re used to a few deals sideswiping you at the end of the quarter that don’t happen,” said Messer, referring to CRM that treats the past as a playbook. “That’s because when you’re looking backward, you don’t see that stuff coming because you’re asking the sales professional what exit she just passed. By using predictive analytics, what we’re doing is we’re telling you about what the most likely outcome is with real-time alerts for opportunities and risks so that you can see the dog that runs in front of your car and swerve or take a shortcut to the destination.”
“When you see this,” added Messer, “you realize why the old world is broken and the enormous potential for sales teams that embrace the future.”