Transforming Big Ideas Into Big Products
What would the Web be like without social media, relational databases or first-rate search engines? When you look at the biggest and most successful Web companies – today the building blocks of our modern online lives – it is hard to imagine them as start-ups, facing challenges to get traction with users, battling with investors over strategy or struggling to prove they can make it outside of their home market. Informilo talks to Yandex, MySQL and a former Twitter executive about lessons learned.
Turning a great idea into a business isn’t easy. Challenges pop up at every turn. But the results can be amazing. Informilo asked three of the scheduled speakers at Dublin Web Summit what they had to do to transform their big ideas into big products. Their stories are both fascinating in their own right and an insight into how to tackle the kind of issues that many entrepreneurs face.
Scaling User Up-Take
Most entrepreneurs struggle to figure out how to get more people to use their product. And Twitter proved to be no exception.
Josh Elman, who describes himself as a “product guy,” was a big part of a sea change at Twitter, helping transform it from a niche service to the important social force it is today. Moving the service from s geeky and difficult to mainstream and essential was no small feat.
Twitter first gained traction among the geek community at the South by Southwest digital festival in Texas in 2007, but it was a while before it attracted a loyal following by the wider net community. “The problem was that people knew what Twitter was, but they didn’t use it themselves,” says Elman, who now works as a venture capitalist for Greylock Partners in Silicon Valley. “Our challenge was to help people understand why Twitter was useful.”
Elman joined Twitter at the end of 2009. At that point he’d been a user for at least two and a half years. “I’d gone through the same challenge as most people with it; I was following some people, but I didn’t really get how to use it every day,” he says. He soon discovered that he was by no means the only one. “Lots of people tried it, didn’t get it, and left. Millions of people had created accounts, but only a fraction were actually using those accounts.”
To solve this issue he and the team studied the data about Twitter’s users. “We realized that people who hit Twitter seven times a month or more were our core users,” says Elman. “The next thing was to look at the amazing set of data we had about our users: we started querying what we could learn about them. We realized that those people followed more than 30 people themselves, which gave us a really interesting ratio: the two-thirds/one-third ratio.”
Most engaged users fit the pattern: two-thirds of the people are followed to gain information while the remaining third are “people you know… people you have bi-directional relationships with,” says Elman.
That knowledge prompted Twitter to build tools to make it easier for new users to find people to follow. Today, for example, if you’re interested in the British soccer team Manchester United, you’ll see in your suggestions not only the players, but also the journalists who cover the team.
Those tools helped drive more engagement and grow user numbers. turning a small, unintuitive service that was used as a chatty back-channel among the technorati into a staple of the web and a key plank of marketing strategies for pretty much every business sector. Just over five years old, Twitter, which is also a vital part of people’s social lives these days, is reported to be heading towards 250 million users.
Staying True to Your Roots
Underpinning every single web property is its data. New techniques to understand and leverage that data emerge all the time, but much of the web wouldn’t have been possible without MySQL, the open-source relational database. Developed by Michael “Monty” Widenius and David Axmark. MySQL is an essential product that continues to underpin many of the services we take for granted, from Google and Facebook to the blogging platform Wordpress. But turning a great product into a business requires investment and management, and sometimes the aims of the businesspeople can be at odds with the founding philosophy.
“I was a typical programming geek,” says Widenius, an open source crusader. “In five years, we grew from three to 15 people. We were very successful, we did a lot of things right, we did a lot of things wrong: we could have created a bigger environment for open source, for example.”
As with so many tech start-ups, MySQL ran into the perennial conflict: how do you return value for the investors while remaining true to your original purpose? “We were mostly technicians – everyone was for open source, we believed in the freedom of it,” says Widenius. “The management team, hired from normal corporations, had very little understanding of open source: they were trying to make money the normal way, not the open source way. We had big conflicts with the management, who wanted us to go more closed-source and make money.”
The challenge was to protect the original vision and keep MySQL open source. Widenius hoped that would be the case when the company was sold to Sun Microsystems in 2008 for $1 billion. “I thought this was the best thing that could happen: they understood development and open source.” However, when Oracle began to hover over the ailing Sun, eventually buying it in 2009, Widenius sprang into action, concerned that his creation, a vital foundation stone of the web, would become increasingly closed in the hands of Oracle.
To that end, Widenius launched MariaDB, the community-developed branch of MySQL. And he launched a new company to ensure the movement stays true to its roots. Monty Program, the company founded by Widenius to develop MariaDB, now has 21 employees, more than half of whom are full-time developers. It’s a return to the philosophy that drove the development of – and vast, global take-up of – MySQL, and an eloquent response to the challenges thrown up by the involvement of a big, commercial entity in the philosophy of a small, open-source entity.
Despite Google’s efforts to ramp up in Russia, Yandex, the country’s most popular search engine, retains around 60% of its home market, now the biggest Internet market in Europe. In 2011 the home grown champion raised $1.4 billion in an initial public offering on Nasdaq. At the time it was the biggest IPO for an Internet company since Google in 2004.
But no company, no matter how successful, can rest on its laurels. Search is changing: it is now about data structures and meaning as well as algorithms and statistics. And mobile is an important part of the picture. Yandex has been buying four to six start-ups a year and is hunting for more in areas like advertising, geolocation and voice processing.
Along with improving and expanding its product lines, there are big opportunities to move into high growth markets. Yandex is hoping to get a leg up on Google by introducing search engine technologies that are particularly well adapted to developing markets. For Yandex, the first step is Turkey, which is worth an estimated $200 million a year and has an estimated 35 million Internet users. But even for a company with proven search engine technology and a huge war chest, the leap from local to international is not an easy one, says Yandex Chief Technology Officer Ilya Segalovich (pictured on Informilo's home page), who co-founded the Russian search engine company with Arkady Volozh.
In general, when the dominant player in a market has a share of between 60% and 70%, the number two should hope to have between 20% and 30%. That’s the goal for Yandex in Turkey, although at present, Yandex has less than one percent of the market, says Segalovich.
“The Turkish market is the first one for Yandex where people do not speak Russian,” he says. Simply establishing a business presence in a new country is a challenge in itself. Beyond hiring local people for marketing, assets and project management, Yandex had to rebuild internal management systems and all internal communications and processes at its headquarters on a new platform and in a new language — English. “It took about a year but now the internal communication system is quite scalable for other non-Russian speaking markets — if we decide to go there,” says Segalovich.
But it’s not just about the administrative framework when you’re building a foreign-language version of a web presence that’s underpinned by the most technical of foundations. A search engine is all about big data management and ranking results. It is all about assessing users’ queries, understanding what they mean, and delivering ranked search, says Segalovich. “The more queries we get, the more ‘meat’ we add to our ranking system, and so the closer our search results can be to what the users actually asked for," he says. “When Yandex came to Turkey we already had quite a few user queries to help us determine what users are really looking for and to help us tune our ranking system.”
But as with any big-data application, the better the analysis of the data that goes in – in this case, user queries – the better the result that comes out. Yandex had to develop several heavy-weight algorithms to be able to assess and rank search results even based on a low stream of queries. “Those algorithms let us deliver relevant results in Turkey even though we are in the early stages and significantly improve search quality for one of the most complicated languages in Europe – Turkish,” says Segalovich. “We keep improving and we keep growing in Turkey as a second search engine thanks to this technology strategy and experience.”
Yandex is also busy setting itself apart from the competition through new services. For example, one of its most important — and popular — differentiators in Turkey is its traffic tracker application. Such technology is lacking in Russia and many developing countries. So, Yandex is applying the real-time traffic technology that it has been pioneering for seven years in Russia to Istanbul and Ankara. While no search provider can challenge Google on a global scale, Yandex is hoping that a more local spin can help it succeed in developing markets. Its approach — to look at what is missing in a market and then develop that – is sound advice for start-ups across the board.