It takes a long time for a large company like IBM Corp. to pass on quality leads to its partners. Lori Cohen used to get one per month, but with the help of IBM’s new AI-buddy SCORE, that number is now eight.
“And of those eight, I would say six or seven are bang on for our company,” said the vice-president of marketing for KPI Digital, an IBM Gold Business Partner.
In its first year of operation, SCORE (the Smarter Cognitive Opportunity Recommendation Engine), has increased lead passing speeds by 50 per cent with a five-point improvement in the win rate on leads passed to business partners. It’s also helped contribute $100 million in additional revenue thanks to the faster sales cycles.
But it wasn’t that long ago when IBM was manually passing leads to its business partners, basing those decisions on limited knowledge about what those partners specialized in and what their target audience was. IBM decided to conduct a study two years ago to understand what worked and what didn’t when it came to lead generation. The results showed that leads passed within 48 hours closed with a 10-point higher win rate than those which took longer to pass. Additionally, less than 25 per cent of leads passed on to partners, it turned out, included relevant information that would help them make a decision.
IBM turned to its research team in Ireland to come up with a solution. Together with IBM’s Chief Analytics Office (CAO) and Chief Information Office (CIO), the research team used Watson AI to create a new search algorithm that eventually became SCORE.
What’s really impressed Cohen, she said, is SCORE’s ability to accept feedback and get better at its job. When SCORE passes on a lead that doesn’t pass the smell test – this could be for various reasons, according to Cohen, while stressing it doesn’t mean the lead is bad, just not the most suitable for KPI Digital – she can tell SCORE why she turned it down, helping the system continuously improve its recommendations. That feedback is stored and the lead is then passed on to another partner perhaps better-suited to follow through on it.
“[IBM] gets to learn about what our strengths are and where our success stories are, as well as how exactly we can benefit customers,” said Cohen. “I’m finding we’re getting many more targeted opportunities.”
Pulling back the curtains and revealing how exactly a machine learning algorithm spits out a decision, or in this case, a lead to a specific partner, isn’t easy. But the minds behind SCORE thought of that as well.
Oznur Alkan, a research scientist at IBM Research, said SCORE considers several things before passing on leads, and it actually reveals this information to channel partners using the service.
“Users want to see the justification behind why the system recommended them,” Alkan told CDN. “The model takes into account several dynamics, like the partner’s current win rate, the experience of the business partner, the deal size, the potential customer’s proximity to the partner, and whether or not the partner performs better with new or existing partners.”
Coming soon to the IBM Partnerworld site is a new “Business Partner Trajectory Model” that will evaluate new business partners entering the ecosystem, according to Michael Fino, channel chief operations officer for IBM’s Partner Ecosystem. This will help new partners, who don’t have a track record within IBM’s partner community, to get recommended by SCORE.
“This new model will look at new business partners and determine if they exhibit characteristics that other successful business partners in the past have shown. And if they have, we are going to increase their ranking so that they will get opportunities from SCORE, too,” explained Fino.
IBM received a Stevie Gold Medal this year for its work with SCORE for “Improving Sales Win Rates with Cognitive Modeling and Process Automation.”
