Telefónica is far from a startup — it’s one of the largest telecommunications providers in the world, and traces its history back more than 90 years.
Despite that, the Madrid-based company, with 123,000 employees, has been embracing the lean startup practices made famous by author and entrepreneur Eric Ries — quickly building prototypes of new ideas developed in conjunction with customers, iterating fast to refine those prototypes, and speedily killing projects that don’t pan out.
“In reality, the situation startups find themselves in when they are testing new ideas is very similar to an innovation team, when they have an idea and they want to make it real in the face of uncertainty,” says Telefónica innovation manager Susana Jurado Apruzzese, left. “The level of uncertainty is very similar.”
The company began experimenting with the lean startup approach in late 2011, and since then has been able to boost its annual number of innovation projects by about 45 percent — without any increase in budget. Projects now progress through the company’s innovation pipeline about 2.6 times faster than before, says Jurado.
Previously, she says, the company’s innovators would generally spend several months turning new ideas into prototypes and only at that point would they begin speaking with actual customers. Often, the prototypes would turn out not to meet actual customer needs, sending everyone back to the drawing board, she says.
“We realized we needed to [test] the market from the very beginning,” she says.
Now, the company sends designers and engineers out to talk to customers about ideas early on — something she says was a “huge cultural change.”
Meeting Customers where They are
For consumer-facing products, they’ll meet with potential customers literally on the street, or on university campuses, or even while they’re waiting in line at a museum. And for business-to-business offerings, they might use company or personal connections to arrange meetings in potential clients’ offices. The goal is to talk to customers in the context where they would actually use the product being discussed, Jurado says.
“This is not the usual way of proceeding in Telefónica,” says Salvador Perez Crespo, right, a senior manager for technology dissemination at the company. “You imagine we have one of those very big laboratories where we can bring customers — [but] that’s not the tool that we use when we’re trying to use lean startup.” Instead, they go to where the customer is.
And when they first meet with customers, they’ll try to avoid discussing the product concept in too much detail, instead focusing on the problems the customers are facing, says Jurado.
“The first thing we do is gather evidence that the problem is real and that it’s painful,” she says. “The kind of questions they ask…[are] ‘what kind of problems do you have’ and ‘how do you solve it now?'”
With one product, Telefónica’s implementation of an industry standard called Mobile Connect that replaces hard-to-remember passwords, developers realized from early interviews with about 50 people that the tool just wasn’t that crucial for the average person on the street, she says.
“Everybody had a solution,” she says. One man “had a piece of paper that he was carrying in his pocket with his username and passwords and he was happy with that.”
The people who weren’t happy with those types of solutions, the company realized, worked in finance.
“They realized that the people in the finance world were having a problem, because these kinds of behaviors were causing security problems for the banks,” she says. It only took two or three weeks to realize they should initially develop a finance-centric application for Mobile Connect, rather than one targeted at the average consumer, she says.
“If they hadn’t done this before doing the prototype, probably we would have had to throw the prototype into the dustbin,” she says.
Don’t Start Building until You Know the Problem Exists
Even after ascertaining what a suitable problem is, and which sort of people are facing it, product developers return to customers for what they call “solution interviews,” showing them simple mockups, sketches, or PowerPoints to make sure the solution they are creating is the correct one, she says.
“Until you have clear evidence that the problem exists, that it’s painful, and that the solution you are going to build is the best solution to that problem, you cannot write a line of code,” Jurado says. “Once you are really sure the solution you’re going to build is the best for that problem, you start building a prototype, but incrementally.”
As they build a prototype, they return frequently to potential customers — those early interviews have usually identified potential early adopters who are happy to help out — to test new versions. They’ll try to lay out concrete hypotheses and metrics for success. For example: a new feature will boost the number of users clicking a particular button by a certain amount. They, then test whether those predictions come true.
“According to the results, you learn, and you start the cycle again,” she says. “If it’s invalidated…maybe you have to make a pivot.”
Much of this testing and developing is done by employees in Telefónica’s R&D division, which has more than 600 employees worldwide.
Teams working on a particular project are at times distributed around the world, and sometimes include employees from other divisions of the company who are assigned to innovation projects after submitting ideas to periodic “innovation calls,” says Jurado.
Getting Business Units Involved Early
And when new products come to maturity, they ultimately have to be transferred to one of the company’s business units for roll-out. That process has evolved over time, says Jurado, and innovation teams now work to involve potential business unit sponsors early on in the process, so that they’re ready to adopt the project when the time comes.
“They have to become sponsors of the project almost from the beginning,” she says.
Recently, the company has also been transferring some innovation team members, with the knowledge and customer connections they’ve accumulated, to the business units to help shepherd the new offerings to market, she says.
“What we’ve realized is it’s like their baby — they want to see their product commercialized,” she says.
Creating a Separate Brand
The innovation teams also work closely with Telefónica’s branding department to create suitable marketing and packaging materials for early-stage products, she says. The branding department has helped create a “Telefónica Lab” look-and-feel for some early products, but some rough prototypes are tested with no brand at all, until the products and services are more solidified. (Note the lab branding on the “Internet of Things” development kit, at right.)
That has enabled the company to test new ideas with the public without compromising its decades-old reputation for reliability, says Perez.
“We have [spent] so many years trying to provide really reliable services,” he says. “For us, it’s something we don’t want to lose.” But the lean startup approach Telefónica has been taking means they don’t have to risk the company’s reputation for reliability in order to run experiments.
For the complete slide presentation, “Lean Elephants: Lean Product Development in a Very Big Organization,” visit our Resource Center.