Salesforce turns to machine learning

For all the advances technology has brought to corporate life, salespeople still rely heavily on guesswork when it comes to wooing customers and closing deals.

|||

San Francisco - For all the advances technology has brought to corporate life, salespeople still rely heavily on guesswork when it comes to wooing customers and closing deals, usually involving late-night conference calls and spreadsheets about client interactions.

A raft of startups such as Clari, Aviso and Highspot are aiming to change that by hawking technology that uses software to analyze and predict deal outcomes and suggest marketing strategies for corporate sales teams, potentially saving time and boosting revenue. For salespeople it can mean answering a short mobile questionnaire in the cab between meetings rather than a brain dump days later when much has been forgotten.

The systems use machine-learning technology to let businesses crunch reams of data that can help them focus time and resources on winnable deals. Larger companies are also taking notice - this week, Salesforce.com unveiled SalesforceIQ, integrating machine-learning capabilities into its main sales-management program to eliminate some of the tedious data entry required to use it.

“These companies have a ton of data, but mostly the data has been used to support human-driven hypothesis,” said Mike Gualtieri, an analyst at Forrester Research. The algorithms “figure out what are the six things that really matter and those six things might be very different than what the humans thought.”

Venture money

While the market is nascent, venture capital firms have poured more than $100 million into young companies that are landing large customers such as VMware and Booking.com, which tout their effectiveness at predicting sales outcomes and focusing marketing efforts. Mobile analytics provider Mixpanel raised $65 million in a single round from Andreessen Horowitz in December to fund a plan to help app makers predict things like which of their free customers will opt to pay and which will leave the app.

For many companies, the current sales-prospecting process takes place on multi-hour conference calls on Sunday nights between the sales force and executives. Participants pore over the past week’s calls and try to plot strategy on each account. They often use spreadsheets and customer-management programs that aren’t designed for making predictions, and can contain information that’s already outdated.

“Companies are at the point where they want to not just look back at the data they have, but use the data to understand what’s to come,” said Stacey Bishop, a partner at Scale Venture Partners, which has funded Aviso. “Lots of our portfolio companies are struggling to spin up an Excel spreadsheet on deals that they need to zero in on, and by the time they do, it’s out of date.”

Acquisition targets

Established companies are also circling, looking for deals that can add capabilities to their existing products. San Francisco-based Salesforce.com spent $340 million to acquire RelateIQ, which combs e-mail, calendar and smartphone-call data to add analytical capabilities into its tools for salespeople. Microsoft Corp. this month bought Seattle-based startup VoloMetrix Inc., which companies like Symantec Inc. use to scan e-mail headers and customer-management software to predict what will make an effective sales rep.

Robert Wahbe, founder of Seattle’s Highspot, held engineering roles in Microsoft’s server software unit before being put in charge of marketing. He found reams of promotional materials, case studies and slide decks - some of them with multiple versions that were only slightly different. There was no information on what worked, or whether clients even looked at the material.

Now, Highspot tells companies like Booking.com which marketing materials get read and predicts which case study is most likely to change the mind of which prospect.

36 spreadsheets

Clari and Aviso were each founded by entrepreneurs flummoxed by the lack of effective sales tools available at their previous startups. One customer Aviso studied had 36 sales managers filling out 36 spreadsheets. The sales operations team would collate them into 10 files to send to the vice presidents. It took several days before they could be turned into one file that the worldwide head of sales used to gin up the sales forecast.

“This is the state of the art in large public companies,” Aviso co-founder KV Rao said.

Rao teamed with an ex-Wall Street quant named Andrew Abrahams, who had switched from serving as JPMorgan’s global head of corporate quantitative research and model oversight to what his LinkedIn page lists as “milking assistant” on his farm in California. They took tools formerly used for analyzing complex financial instruments and revamped them to assess the risk to each deal in a customer’s sales pipeline. Aviso now has 32 customers including Splunk and Nutanix, Rao said.

On the fly

Clari, based in Mountain View, California, has a mobile app with a few queries that salespeople can fill out on the fly between meetings, designed to cut down on the manual data entry that traditional customer-relationship management programs require. Its machine-learning algorithms analyze historical information on deals won or lost, as well as e-mail and calendar data showing who participated in meetings and exchanged messages. All that gets combined into a calculation of which deals are progressing toward a win, said Clari co-founder Andy Byrne.

Using Clari let VMware’s networking unit, formerly Nicira Inc., do away with Sunday night forecast calls that ran through every single customer, said John Jendricks, VMware’s vice president of business development and operations. Instead VMware holds a short catch-up call on Monday and managers have current information on customers at all times.

Mobile-app users

Mixpanel, meantime, is using the Andreessen funding to try to bring some of these same capabilities to vendors that traffic in consumer users rather than business accounts, helping to predict which mobile-app users are likely to switch to paid versions, for example.

With all the startups and venture money, the space is getting a little crowded, Forrester’s Gualtieri said. It’s also hard to predict winners. With established companies like Oracle Corp. and Salesforce pushing in, not all the companies will survive, he said.

“Some of these startups will get acquired and some will disappear,” he said.

Bloomberg Beta, part of Bloomberg News parent company Bloomberg, is an investor in Aviso. Bloomberg also is an investor in Andreessen Horowitz.

BLOOMBERG