ML-Powered Digital Move Quote

Getting a move quote was too heavy, too early

For interstate moves, getting a quote was traditionally a deep-funnel process. A customer would connect with a sales rep, share initial move details, complete a virtual survey, and wait for a formal quote to be assembled based on home contents, services needed, and move complexity.

That process worked for highly committed buyers, but it created friction much earlier in the funnel. A large segment of customers simply wanted a fast, credible ballpark quote so they could compare options and understand whether a white-glove move was even in range for them.

The gap was speed and confidence

Leads were being lost because they could not get enough pricing clarity early enough. The quote came too late, after too much effort, and by then abandonment had already started to compound.

This was not just a pricing problem. It was also a product problem. Customers needed a faster path to information, and the business needed a better way to capture and qualify demand without forcing every lead through a high-touch sales motion too early.

A faster way to price the move

UniGroup had decades of moving data across the country, including detailed surveys, services purchased, and pricing outcomes. We used that foundation to train a machine learning model that could generate ballpark quote ranges earlier in the funnel.

The model incorporated variables like ZIP code, home size, services requested, building floors, room count, seasonality, move timing, and other supporting data signals. The goal was not to replace the final quote. It was to give customers a credible pricing range early enough to help them make a decision.

That shift did two things at once. It gave leads useful information without requiring immediate sales involvement, and it gave the business a better way to capture and nurture interest through more personalized follow-up and sales enablement.

Just as importantly, we had to decide when the estimate was good enough to show, how to present a range honestly, and how to balance customer usefulness with business risk.

The impact

The approach proved itself quickly.

Non-converting leads dropped by 50%, and conversion improved by 30%. More importantly, the experience better matched how customers actually shop: they want useful pricing guidance early, not only after they have already invested time in the full quoting process.

This was a strong example of using machine learning in a practical way. Not for novelty, but to reduce friction, improve qualification, and make a complex service easier to buy.