FILE - FEBRUARY 19, 2013. (David McNew/GETTY IMAGES)

A year and a half ago, Matt Thier, co-founder of online-only clothing store BetaBrand, thought he noticed more people in San Francisco wearing bold, patterned socks than before.

Thier had been thinking about expanding BetaBrand’s sock line beyond the striped socks it already sold, but he wasn’t sure if there was enough interest, or if BetaBrand should continue investing in socks at all.

So Thier decided to do some market research, not by surveying shoppers, but by analyzing Internet searches.

He used Google Trends, a free service showing the volume, popularity over time, and geographic concentration of particular search terms. Google Trends is one of several search-term analysis platforms currently available, including subscription-based services like Compete, which can be a few hundred dollars a month.

Thier found that people all over the United States — not just San Francisco, where BetaBrand is based — were indeed searching Google for “bold socks,” “patterned socks,” “cool socks,” and “crazy socks,” and these terms had increased in popularity over the past several months.

With this evidence, BetaBrand launched a new line of patterned socks as well as a promotional campaign called“Sock Insurance” — customers are entitled to two free replacement socks for up to a year for each pair purchased.

The company was pleased with the results. BetaBrand had initially ordered up 400 pairs of socks at a cost of about $4 a pair — it sold the socks for $13 and soon was ordering more.

So far, about 4,000 pairs have been sold; about 30 customers have claimed replacement socks as part of the insurance program.

Before it used analytic tools like Google Trends, the BetaBrand team used to rely more heavily on personal hunches.

“Our products tend to be responses to things we see going around,” Thier said.

In 2008, the start-up released one of its most popular items — “Bike to Work” pants, or dress-pants with reflective-inner lining designed for bikers — because Thier and his team had “noticed people were biking to work, and there was nothing on the market for bike-to-work pants.”

Search term analysis tends to be more useful to companies that can react to changes in trends quickly, said Francis Skipper, executive vice president at 451 Marketing, a digital marketing firm based in Boston.

“If you’re trying to gauge interest, by the time you see it as a trend, you’re going to be second or third to market,” he said, noting that by the time search terms show a noticeable spike in popularity, businesses may be a couple of months behind if they aren’t already selling the product. “From a smart marketing standpoint, it’s too difficult, or too risky, for brand new forecasts.”

Indeed, Thier said, search-word analysis is only one tool. The company also relies on customer feedback and other research, including its own hunches.

The process doesn’t always yield success. Several months ago, Thier was looking for a way to update one of BetaBrand’s most popular product lines, called “Discowear,” featuring clothing made of a metallic, highly reflective fabric. He’d noticed a growing cultural interest in mixed martial arts — it was on television more, more people were talking about it — and he confirmed, using Google Trends, that more people were searching for the sport.

So BetaBrand released what Thier thought would capi­tal­ize on the cultural trend — a “fairly ridiculous item” called the “Disco Kimono,” a metallic-silver robe inspired by martial arts uniforms.

BetaBrand produced about 150 kimonos, but they sold slowly, and the team decided not to order additional units.