Gusto provides a cloud-based payroll, benefits, and human resource management solution for businesses.
Gusto offers health benefits and workers’ compensation benefits that include medical insurance, commuter benefits, and 401(k) contributions as well as workers’ compensation insurance plans, time tracking and integration services, and access to employees for processing payroll from web-enabled devices such as smartphones and tablets. It also provides access to employees and contractors to browse their previous pay stubs, review their payroll forms, verify their personal details, and statistics about their pay, tax payments and federal forms filing services, and security and support services
Gusto serves startups, coffee shops, doctors, creative agencies, lawyers, and boutiques. Edward Kim, Joshua Reeves, and Tomer London founded ZenPayRoll in 2011 that became Gusto, with its headquarters in San Francisco in California with an additional office in Denver in Colorado.
Gusto offers a self-serve provisioning experience as well as a high touch sales model. Because Gusto generates a high volume of leads and serves SMBs, sales efficiency is paramount to mantain CAC in an acceptable range.
Gusto assumed their inbound was made up of very high quality leads that need sales, others that could self-convert and very low quality leads that will never convert. The marketing team wanted to ensure reps were spending their time calling the high quality leads that could be impacted by the touch.
MadKudu connected to Salesforce, Marketo, Segment and started analyzing historical cohort data to train a model that identified the highest quality leads. As soon as a lead would get created in Salesforce, MadKudu would assign a score from 0 to 100.
The experiment we ran was to compare the conversion rate difference of a given leads cohort between when called vs. not called. For illustration, if a certain cohort converted at 95% when called and not called, it would belong to the self-serve segment because reps calling led to little impact in conversion rate. To reduce complexity leads were bucketed in "MK Good" and "MK Bad" groups based on their MadKudu score. Each group was split in a test and control group. The control group leads were called but the test group leads were not called, the conversion rates were measured after 60 days of testing.
The MK Bad group showed no statistical lift in response to reps calling whereas the MK Good group displayed 15% lift at 98% confidence.
This was a clear demonstration that Gusto should continue calling any lead that is scored "Good" and prioritize their efforts against that group while maintaining a strong self-serve option for the "Bad" leads.