Table of Contents

I joined Advent International as just a part-time 10 hour/week contract for a few months. Despite the lower-end time commitment, I managed to provide a lot of insights for existing portfolio companies as well as providing the data science team data to advise on companies Advent was considering to purchase.

Highlights

• Minimized cannibalization based on brick-and-mortar store locations
• Designed software to estimate the number of customers on potential investment opportunities
• Created programs to generate insurance quotes to better understand the algorithms providing the quotes

Cannibalization Analysis

For this project, I was given historical records of each brick-and-mortar store for this company. They’re a package delivery company, which has physical location to visit. A large challenge was determining if the dips in physical store throughput was caused by the holidays or cannibalization as a result of a store being too close.

Unfortunately, we were working with extremely limited data only going back one year so it was difficult to determine the exact cause. Despite this challenge, using general trends of package delivery across the holiday seasons we were able to estimate which brick-and-mortar stores were being cannibalized and identified new locations to build to minimize cannibalization (albeit with a higher degree of uncertainty than I would’ve liked).

SaaS Customer Estimation

Most SaaS platforms seem to use the general domain of your_company.sass_domain.com, this has advantages into giving your customers a customized dashboard and experience. We realized that one of the companies Advent was considering to invest into was using this same structure.

We extracted data from LinkedIn and a few other sources to find the names of companies that may or may not be customers of the platform. I then identified an API endpoint the client made to check if the subdomain was valid and we utilized that endpoint to determine if the names of these companies (or variations of it) had existing dashboards on the platform. If a dashboard existed, it was a strong indicator that the company was at least a former customer of the platform.

This data was then used to see which customers were currently using the platform. We analyzed the companies that we found were using it to determine how large they were and estimate how much per month they were likely spending on the platform to estimate a rough MRR for the SaaS platform.

Insurance Quotes Generator

The last project was much shorter than the other ones. Advent was interested in an insurance company. One critical part of this was to figure out roughly how they were calculating insurance quotes and compare this to other more main-stream insurance providers.

I wrote a program that would find quotes for various insurance companies by using a ton of different profiles of real people. These attributes are the typical fields you’d fill out in insurance forms like: address, age, gender, etc. There were around 300 of these profiles to test on a handful of insurance providers to gain insights into the differences in how they computed the insurance premium.

Overall, I feel like I learned a lot while at Advent International, and I did find the investment world to be interesting. Although, because my time commitment was only 10 hours a week I didn’t get to be apart of the entire process of purchasing a business as much as I’d like to.

Back to Home