Close

Dancing with elephants

A cautionary tale of one of our most promising Web 2.0 startups

A cautionary tale of one of our most promising Web 2.0 startups

Juha Saarinen

[Technology]

You’ve got a hot new idea for a product or service that will slot neatly into an existing infrastructure. You can create it quickly and build on someone else’s audience. But are you living on borrowed time? Will your carefully developed and devised offering survive, if it piggybacks on someone else’s service and is at the mercy of business decisions that you have no influence over?

Those are some of the questions a startup company probably should consider, but few do. I’ve lost count of the number of pitches received about neat little services that use the back-end of a big corporation offering. There seems to be a never-ending stream of people working on a great app that uses Google-Something or Yahoo-Wotsit for data or transport.

And how about all those SMS-based services that are popping up like mushrooms in the autumn? You know, the ones where a provider like Vodafone takes two-thirds or four-fifths of the revenue each text message brings in. You’re entirely dependent on access to the provider’s network and it’s never on your terms.

Is this a good way to build a business? The simple answer, of course, is ‘no’, but the picture is much less black-and-white than that. Case in point: the rise, fumble and stumble of Christchurch startup Eurekster. Co-founder Grant Ryan concedes the company fell under Google’s elephantine feet, but there’s more to that story.

Eurekster first popped up on the radar in 2004, when it launched what was termed ‘personalised social search’. The search engines are called Swickis, and the idea is that they learn from a community of enthusiasts and experts.

That way, results are customisable, so that users searching for particular information could be directed to sites within a cluster covering a certain topic of information, or selling specific products.

For a while, things went from good to better for Eurekster. Media, tech and basketball entrepreneur Mark Cuban boosted Eurekster, calling it “the best sales idea I have seen in a long time”. That was in 2006, the same year Eurekster was named a Business 2.0 Next Net 25 company, received a Red Herring 100 North America award, and BusinessWeek reported that Microsoft was in talks to buy it.

With the positive buzz surrounding the company, it was only a matter of time before US investors woke up to Eurekster and stepped in. With US$5.5 million of venture money and offices in Christchurch and San Francisco, things were looking very good indeed for Eurekster.

It wasn’t to last. Late last year the Christchurch office closed; then in May eurekster.com went offline and the Swickis stopped working. Upset comments from people who had spent a large amount of time customising Swickis for their businesses swamped Web 2.0 online forums. (As Idealog goes to press, Eurekster.com and the Swickis are back up again, run from the US.)

What happened? Well, initially, Eurekster’s Swickis were successful because of Google. And ironically enough, the wheels fell off Eurekster because of Google. Ryan says changes to Google’s mighty algorithm meant many Swickis dropped out of the search giant’s index. “With 80 percent of users finding things on the Internet via Google, if you’re not in the index, you don’t exist,” he says.

When Eurekster started, Google wasn’t anywhere near as dominant as it is now. It was difficult to envisage a future where 80 percent of your market was in the control of another company, and at the whim of arbitrary business decisions.

Ryan isn’t bitter about Google doing what it did. The changes to its search algorithm were likely driven by a desire to improve the service for users and to make life harder for spammers, he says, and not malicious. Unfortunately, Eurekster’s Swickis got caught up in that change and suddenly Swickis were a lot less attractive to website owners.

Despite Eurekster’s failure, Ryan doesn’t believe it means you should necessarily shy away from large partners if it’s the fastest way to get to market. Look at what’s on the table, he says: plenty of people have been very successful developing apps for Facebook for example, and a willing and communicative large partner can offer plenty of opportunities for great ideas.

How do you minimise the risk then? Ryan says it’s tricky to get the balance between risk and certainty right, but open APIs (programming interfaces used to share data) are key to avoid having your toes mashed while waltzing with the giants.

An example is Google Maps. The search engine behemoth is unlikely to turn off access or make radical changes to the Google Maps APIs, as it wants them used by as many websites as possible.

And building on top of a large company’s product makes you a more attractive acquisition, too. Ryan has some advice he’d like to share: if a Google-sized company puts what seems like a fair amount of cash on the table and offers to buy you, for God’s sake, take it.

He speaks from bitter experience, for he says this is what happened to Eurekster last year. Ryan—who knows a bit about these things, having earlier sold GlobalBrain to NBC for US$32 million—thought Google’s offer was good and wanted to sell, but the US investors vetoed the deal, taking a gamble on continuing the business and then perhaps getting more money at the end. That gamble failed, leaving the investors to ponder a possible loss of their investment in Eurekster. (Ryan won’t be pondering with them; he left Eurekster “some months ago”, and says his only involvement now is helping out with some intellectual property issues.)

The final piece of advice from Ryan is not to get discouraged. Yes, he’s disappointed that the punt didn’t succeed, but says the love of startups is still strong in him and he intends to continue being an entrepreneur and inventor.

Ryan is reluctant to talk about his next venture, but says it won’t be software. “I’m a mechanical engineer by training,” he hints, perhaps with a note of relief in his voice that there’s no equivalent to Google on that particular dance floor.