Chris Barton, founder and creator of Shazam has admitted that his team nearly gave up on the company in the early days of its establishment.
At Tech Show London on Wednesday, Barton recounted challenges faced by his team, including co-founder and chief scientist Avery Wang, who were often on the verge of abandoning the tech start-up after months of difficulties building the music recognition technology.
“Simply having an original idea is not enough,” Barton said, noting that the development of Shazam also required a mix of effort, creative persistence and emotional commitment to achieve success.
“We almost gave up, multiple times,” he added. “There were times where Avery would come and say, ‘I am sorry’ but this startup is not going to work, because he couldn’t come up with the right invention.”
In June 2000, Avery finally “cracked the code” and developed the breakthrough technology that would become Shazam’s core AI innovation.
“The technology used AI pattern recognition, three dimensional graphs of energy, graphs against time and frequencies and combinatorial hashing for fast search, and was considered the first mass consumer AI,” Barton said.
The founder then explained how the start-up developed over the years what he called “creative persistence” to solve problems by sidestepping traditional barriers. He shared examples of additional obstacles the start-up faced in the early 2000s that required creative action, in a world where digital databases did not exist yet.
“There was no Spotify, there was no iTunes, there was no Apple Music, there were not digital music databases that we could get to,” he explained. “So we had to create our own digital music database.”
According to Barton, Shazam’s first database was built manually through a partnership with a CD distributor, hiring a few young workers who worked 24 hours a day to pick CDs from racks and transfer data to a custom-built software.
The application was first launched in 2002 with a unique four-digit phone number on Nokia, where users had to dial the number and hold the phone toward music to identify a song.
“This initial model had a lot of friction, which made adoption difficult until the App Store launched in 2008,” Barton said.
“Users had to know the specific phone number; it wasn’t as simple as pressing a button on an app,” Barton recalled.
He added that other problems included cost friction, as the business model charged 50 pence per song recognition, and technological limitations, including no options for downloading music, noise cancellation, audio filtering, seed variation.
“Sometimes pebbles that are actually boulders, and small points of friction that seem insignificant but can actually be major obstacles to user adoption,” he explained.
Despite early struggles, including six years of financial instability, the company came up with new features such as synchronised lyrics and offline Shazam to differentiate larger competitors including Phillips, Sony and Google.
“The innovative approach has led to Shazam now having 300 million monthly active users and an acquisition completed by Apple in 2008,” he said.
