Rank prospects based on buying stage, interest level & worth
Problem:
Low lead conversion.
Sales rep do not know where to focus efforts on.
Do not know whom to target from the external database of millions of Americans.
Data Science Challenges:
Many data points were not linked to IDs. The herculean task to join tables and at many places, we had to do some guesswork.
Later, the machine learning models were doing as per expectations. We had to use man latest models and build ensembles out of them.
Approach/Solution :
Ranked and gave numeric values to the leads. Higher the score, the better the chances of leads converting into sales. This numeric value took into account both probability of buying into account and how much they will potentially spend in this company.
Automatic lead scoring on future leads the moment they come into the database. In other words, we have automated the scoring on new people.
Gave a cutoff score, below which wasn't worth pursuing a lead. We did a lot of if-else scenarios, took costs into considerations and came up with this number.
Results:
The conversion went up from 5% to 8%.
Fewer temporary Sales Representatives hired to manage the workload of leads.
On average cost saving of 70 cents per unqualified lead.
$21,000 more revenue every month due to focus on high value leads.
