Online Extra: Q&A with's John Browett

How has this British online grocer become successful? Says its chief: We never fell into the group-think mentality

John Browett, the 37-year-old chief executive of online grocer, keeps his cards close to his chest. He's more than happy to discuss the success of his company, a unit of Britain's No.1 supermarket chain Tesco PLC. And he doesn't mince words when explaining why other online grocers, such as Silicon Valley's Webvan, have failed.

However, when it comes to divulging the secrets of's formula, Browett is downright cagey. He will allow, though, that reaching today's success was a long, hard slog that had little to do with flash and a lot to do with patient trial and error. Browett spoke with BusinessWeek e.Biz's European regional correspondent Andy Reinhardt in early September. Here are edited excerpts of their conversation:

Q: Give us the lowdown on how is doing, by the numbers.


We did 237 million [$342 million] in the year ended February, 2001, and we're now at an annualized rate of more than 300 million [$435 million. On Sept. 17, Tesco said's first-half sales were 147 million [$213.2 million] and that the dot-com unit would be profitable by yearend.] We lost a total of 9.5 million [$13.7 million] last year, due to the cost of rolling out stores and starting five different nonfood businesses -- electronics, baby and toddler, fashion, entertainment, and wine deals. But the grocery business is already profitable. Over the long run, there's no reason it can't be as profitable as retail stores, which now have an operating margin of 5.7%. We're not there yet.

Q: What makes this any different from grocery delivery boys in the 1950s?


There's no real difference. All we are is the electronic butcher's boy. But the difference is that in the 1950s, supermarkets had gross margins of 50%, and now they're half that. They could afford home delivery back then. Margins have been squeezed and squeezed and squeezed. At the same time, service has become more expensive. We're using technology to deliver a service. The Net allows you to capture and process [grocery] orders in a very efficient manner and, as a consequence, lets you deliver them.

Q: What accounts for the success of


We started in 1996. We did it as a skunkworks, with very limited resources. There were just six middle managers, with a direct line into the CEO, who were told to go away and make it work. They spent 2 1/2 years in 12 stores just refining the process. Like a traditional business, you start small and grow. So, we spent practically nothing, less than 1 million [$1.45 million]. Only then did we roll it out to 100 stores, then 240. That's when we started spending real money.

Q: How does that compare with Webvan's shoot-for-the-moon approach, where it spent $1.2 billion in two years and ended up in bankruptcy?


You have to make the cost structure work and then roll it out. You can't make a run for revenues and then work out the cost structure later. We were never caught up in the maelstrom. The point is always to have revenues ahead of costs. We will have spent 40 million [$58 million] on groceries and 60 million [$87 million] in total on all our online stuff. Now, the nonfood businesses can work off the foundation of food. We launched our baby and toddler site for 500,000 [$725,000], including promotion.

Q: Analysts questioned your decision to pick groceries off the shelves of your supermarkets, rather than build dedicated warehouses. Do you think you have the last laugh now?


We haven't gotten to the stage of the last laugh. We were clearly out on a limb when the hype was at its peak. We had analysts telling us we were doing the wrong thing, people from McKinsey, Andersen, PricewaterhouseCoopers. So we went to visit everybody -- Webvan, Homeruns. We looked at them in detail to make sure we were right. Beneath the hype, nobody was sure these things would work. We never fell into the group-think mentality. It's in our character to go away and do our calculations very carefully. We've been a bit lucky, but we've also been right.

Q: Some critics say's financial results are inflated because Tesco PLC isn't charging you enough overhead. How do you respond?


Some analysts challenge whether we're allocating costs properly. We are. We do a very careful cost analysis. There's no point in fooling yourself. You have to actually make money. All [our new business initiatives] have to make sense in their own rights. We don't want to do cross-subsidizing.

Q: But don't you enjoy some help from Tesco that you're not paying for?


Of course. There are enormous benefits to being a part of the Tesco infrastructure, like volume-buying and the Club Card [affinity card-holder] magazine. It's the largest single mailing in Europe -- it goes out quarterly to 10 million households. We get ads without having to pay for them. The same goes for billboards and banners in the stores. There's opportunity cost to Tesco, but no cash cost to us.

Q: Now that you've shown online groceries can be profitable, aren't you worried rivals will catch up?


It's easy to copy somebody else's ideas and put them in your store. But what most of our competitors don't understand is that this looks simple on the surface, but the detail underlying it has turned out to be very, very hard. That ranges from building a transactional Web site that can handle the volume, to the in-store picking process, to making the deliveries efficiently. Lots of people are delivering services via the Internet, but the question is whether they can do it at our scale. We had frankly expected that by this stage, somebody else would be at our scale and making money from it, but it's not so.

Q: What makes this so complex?


It's the number of items. doesn't have to maintain such a big database for each customer. What we have to do is mind-blowing -- our database is hundreds of times bigger, very complex. We maintain data by customer, by individual product. And then it all has to be integrated with the back-end systems. We pick 4.5 million items per week -- more items that any other online business in the world.

Q: Why did Webvan fail?


The fundamental problem was that they overestimated demand in the early years. Second, they rolled out a business before they had a proven business model. Third, they allowed their agenda to be driven by capital markets instead of by business fundamentals. Perhaps if 10% of all grocery volumes were going through the Net, Webvan might have worked. But if it was only 0.1%, the delivery economics didn't work. They rolled it out before the market was ready.

Q: Were there other problems?


There was too much distance between drops. The warehouses were designed to serve a large area to get enough volume to feed them. By comparison, we can economically serve an area of 100,000 people, not several million. We can get to breakeven with very low volumes. With their model, it's possible they might never have reached a high enough volume to make warehouses cost-effective. Plus, it wasn't just the wrong business model, they also made some errors of execution. In trying to drive toward very low picking costs, for instance, they often made poor choices in machinery. Webvan was technology gone mad.

Q: Critics worry there's a ceiling on your growth capacity.


We think we can go as high as 2 billion to 3 billion [$2.9 billion to $4.35 billion] in sales using store-based picking and packing. And that's assuming no new stores and no significant productivity improvements -- both of which are likely and would raise the number. Maybe someday we'll have a hybrid model with warehouses for dense areas like London, but the traffic just kills you. Plus, where would we put them? In any case, smaller places like Inverness, Plymouth, or Peterborough could never be served from warehouses.

Q: What can you tell us about your deal with Safeway in the U.S. to roll out delivery there this fall?


We knew we had something that worked -- one investor said we had "cracked the code." We went to visit all the U.S. retailers, because the U.S. is the best and most attractive market, and then a number of retailers came to see us. The great thing about Safeway is that it's an incredibly high-quality business. But we have no comment on opening dates, locations, rollout schedule, and so on. We worry about a technology glitch that could set us back three months, so we won't even say what our goal is.

Q: Why are you expanding into South Korea?


We already have seven stores there, with 11 more under way. We aim to launch in the first quarter of 2002. The fact that Korea has the highest residential broadband penetration was merely how we explained it to the press. Most reporters don't understand how sophisticated Korea is. But that's not really why we chose to expand there. We did it because we have stores there, and it's a developed market.

Q: What else does the future hold?


We've got a chance to become the leading "last-mile" delivery service in Britain because we're taking it very incrementally. A good example is our new Wine Warehouse business. It takes advantage of Brackmills, our wine-distribution center in North Hamptonshire. We send whole cases to Tesco store loading docks and then send them back out to customers in vans without opening the boxes.

You couldn't do that business stand-alone, but we can leverage off our existing warehouse and trunking to provide the last-mile delivery. Plus, we can offer much more information about wine than you ever could in a store. The store experience is always about trying to edit down the amount of info to a minimum. Online, you can provide much more to customers who want it.

    Before it's here, it's on the Bloomberg Terminal.