Putting data into the sales process with InsideSales
InsideSales is one of an increasing number of players that aims to offer sales teams tools to make them more efficient and effective.
In its case, InsideSales came about through the post-graduate research of co-founder Dave Elkington. As Elkington studied artificial intelligence, he soon came to realize that A.I. has existed for decades — the math that is behind A.I. was used back in the mid-20th century by researchers at companies such as IBM.
What is different today — and what gives the power to the disruptive companies to undermine their more conservative competitors — is the access to data. As Elkington sees it, Netflix’s ability to put Blockbuster out of business was a direct result of Netflix’s intentional strategy to amass information about its customers and, in doing so, to give its own predictive algorithms the best possible source data to tune its suggestions.
The quantity and quality of the data available today is what is different.
This realization led Elkington to focus his thesis on the idea of an A.I. offering that uses mass anonymized data to identify patterns and linkages. This research was used as the basis to found InsideSales and to approach the company from the get-go with the intention to amass customer data, anonymize it and train its algorithms to be more and more effective.
This topic was the focus of Elkington’s keynote at InsideSales’ recent Accelerate conference (disclosure: InsideSales covered my travel and expenses to attend the event).
In his keynote, Elkington took a sideways stab at his competitors, explaining that it isn’t the core predictive algorithms that the company uses that makes it different. It is the huge amount of anonymized data that the company has, and which it uses to continually tune and improve its predictive algorithms.
To illustrate this point, Elkington told the story of one of the world’s largest SaaS companies that gave InsideSales all of its 2015 sales lead data to run a trial upon. InsideSales ingested the first six months’ worth of data and used it to train its algorithms, whereupon the company in question asked InsideSales to predict sales and closed deals for the next six months. The trained and optimized algorithms did a near-perfect job of delivering the required results.
Elkington announced the next stage in this journey: the creation of InsideSales’ playbooks — data-driven tools that give salespeople direction about who to contact, what channel to contact them through, when to contact them and how to approach the conversation. And the results that speaker after speaker experienced were impressive. From varied organization such as Dyson and CenturyLink, InsideSales and its playbook improved salespeople’s outcomes as well as their efficiency.
I wanted to dive into Elkington’s thoughts about A.I. more generally and with a combination of conversations onsite and via follow-up emails.
From the get-go, Elkington bemoans the fact that while we all use A.I. many times a day in our consumer lives (whether it’s a Netflix recommendation, a Fitbit suggested activity or an Uber dispatch), the reality is that very few workplaces actually have anything even remotely resembling A.I. within them. The closest most get to A.I. is using manually entered rules. The enterprise market is, according to Elkington, awash with confused buzzwords.
« What I’m seeing in the marketplace is companies interchanging the terms ‘predictive analytics,’ ‘A.I.,’ ‘big data’ and ‘machine learning,' » he said. « This is a mistake and these words describe very different things. ‘Artificial Intelligence’ is a broad term within which machine learning is a distinct subset — machine learning is simply the set of algorithms that a particular organization uses to automatically give an answer to a particular problem.
« It’s essentially math that gives an answer to a question, » he continued. « But the reality is that to make A.I. really work, you need mass quantities of data in order to tune the particular algorithms. The reason Amazon has been so good at predictive intelligence is that even as far back as 2005 they were doing close to 100 million unique customer transactions each month — what enterprise organization is processing that many transactions on its own? »
Which is where, of course, InsideSales comes in. By amassing the data across all of its customers, it is able to produce far more accurate algorithms than individual companies would be able to were they working on their data in isolation. InsideSales has, somewhat uniquely in this space, managed to convince its customers that using its data is net beneficial, and no real organizational risk. No mean feat!
InsideSales’s platform, Neuralytics (think « neural networks » plus « analytics »), is the engine that ingests and mines all that data InsideSales holds. In doing so, Elkington has created an immensely powerful web of relationships and behaviors.
« The amount of data in a platform is secondary to how that data is stored and normalized, » he said. « It is the relationship between all the data points which is really of interest. We’ve begun to put a service on top of our data and enabled organizations to use it as their own predictive cloud. One cloud infrastructure company is using our platform to better predict their utilization — they’ve actually increased the accuracy of their predictions from 30% to 70%. Another company, a global entertainment operation, has used our platform to create a more personalized service for their VIPs. »
This is interesting and a trend that I predicted a couple of years ago — that these predictive analytics players would extend their solutions from being vertically specific to being broad horizontal ones. My theory went that once a vendor had built up sufficient expertise and data in a particular area (say, for example, scoring sales leads), it would then be able to extend its platform to be a broadly applicable analytics engine and, in doing so, create afar greater addressable market than before.
Another aspect of this is, of course, that a broad analytics platform creates massive opportunity for acquisition and application to a large enterprise vendor’s product set.
An intriguing thought would be if Salesforce, for example, would acquire InsideSales for the robustness of its analytics engine. After all, we’ve seen big acquisitions of companies with huge data and analytics capabilities before — think LinkedIn by Microsoft or The Weather Company by IBM. Elkington admits that InsideSales’ platform can be more broadly applied.
« Neuralytics is designed to be generally applicable to any data problem. That was, after all, the topic of my thesis — not on building the best sales analytics product, but on building the best analytical platform period, » he said.
InsideSales seems to have done a very good job of differentiating itself from all the other players, having executed its vertically specific opportunity very well and having now embarked on a far more ambitious and broader play.
It would seem that there is every opportunity for this company to be an important player in enterprise A.I. into the future.
InnoValeur | Data Science | Smart Data | Machine Learning | AI