# Why Dell.com (was) More Enterprise 2.0 Than Dell IdeaStorm

In my keynote last week at Web 2.0 Expo New York, I made the comment that, cool as Dell Ideastorm is, the fundamental supply-chain approach behind dell.com is actually a better example of how Web 2.0 applies to the enterprise. I also made the provocative assertion that WalMart is a Web 2.0 company (or at least a model of how Web 2.0 principles apply to the enterprise.)

Based on questions I’ve heard since, I thought I should explain further.

Many people have been seduced into the ideas that Web 2.0 is all about explicit collaboration, contribution, and “the wisdom of crowds.” So, for example, on his Web 2.0 Watch List, Seth Godin wrote: “For our purposes, my definition is that most of these companies are, as the wikipedia says, sites that ‘let people collaborate and share information online in a new way.’ So,” he says, “Google doesn’t make the cut, because most of their traffic comes to their search engine.”

Now, I will say categorically that any Web 2.0 definition that excludes the Google search engine is broken. But it’s broken in an instructive way, one that shows what the problem is.

Web 2.0 is ultimately about understanding the rules of business in the network era. I define Web 2.0 as the design of systems that harness network effects to get better the more people use them, or more colloquially, as “harnessing collective intelligence.” This includes explicit network-enabled collaboration, to be sure, but it should encompass every way that people connected to a network create synergistic effects. So let’s take Google:

• The very act of creating a search engine via spidering relies on a user-contributed network. Where does Google get its raw material except from us? When I publish this blog post, I’m contributing to Google (and to every other search engine.)

• Google’s PageRank algorithm extracts an additional level of implied metadata contributed by links. All spiders follow links to discover new content; PageRank taught us that some links matter more than others. The engine became smarter by understanding a bit more about what people were contributing even when they didn’t know they were doing so. Many a new breakthrough in Web 2.0 (e.g. Facebook’s social graph) comes from making implicit contribution explicit in some way, gathering its benefit and amplifying it.

• Google’s real-time ad auction is the heart of their economic engine. Their stroke of genius, which gave them their seemingly insuperable lead over the other search engines in ad monetization, was to understand that selling the top ad position to the highest bidder was actually leaving money on the table. Given that advertisers only pay for clicks, Google realized that if they could project the likely click-rate on an ad, they could sell top position to the best combination of price and click rate. A $5 cpc ad clicked on 3x as often as a$10 cpc ad will make \$15. By instrumenting, measuring, and responding to the click rate, Google made ad auctions smarter – through harnessing implicit user contribution.

• Google also measures click-through behavior, surfing habits, and everything else they can get their hands on that will help them improve search results or ad performance. All of this is fed back into a real-time loop that is constantly trying to automate and optimize the user experience. Sounds a lot like the human brain, eh? (At least one with lots of learning plasticity.)

If there is only one thing that enterprises ought to learn about Web 2.0, it’s this one: building information systems that allow you to adjust in real time based on interaction with your customers is the true mark of a networked enterprise.

If you understand the following three things, you know everything you need to know about Web 2.0 and the enterprise:

1. Harvest every bit of user contribution, not just the explicit. Your business has thousands of touch points with customers. When they buy from you, they contribute data as well as money. When your suppliers increase their prices, or change their delivery times, they contribute data to you. When you advertise, and people respond (or don’t), they contribute to you. When you introduce a new product, when you do something your customers love, or hate, and people talk about it, they contribute.

Your data is one of your most critical business assets. Are you doing everything you can to wrest competitive advantage from it? I’ll remind you again: PageRank and the real time Adwords auction were both hidden in plain sight. Understanding what data you have, and what meaning you can extract from it, is the holy grail of Web 2.0.

2. The era of IT as a back-office function is over. It’s no longer good enough to gather data and analyze it, then propose and adjust strategies over the next budget cycle. You must infuse your organization with IT, so that, like Walmart, your supply chain responds every time a customer rings up an item at the cash register. This is how Walmart is like Google. No, not the website, but the live enterprise, which learns and responds.

That’s why in my enterprise 2.0 talks, I usually end by saying “turn your IT department inside out – or wait for some innovative startup to do it for you.” Banks could be building something like Wesabe’s Value Engine and tips feature, which extracts collective intelligence from credit card data; phone companies could be doing something like Skydeck‘s extraction of your social network from your phone bill. In fact, they’d be in a way better position to build integrated services against this data than startups that are having to first extract the data from corporate databases one customer at a time!

3. Web 2.0 thrives on network effects (also known as virtuous circles): data begetting more data, services getting better in such a way that they are used more often, until you are so far ahead of the next guy that he can’t catch up. That network effect is enhanced by letting other people use and build on your data, not by keeping it private. What we’ve seen is that the first company to create network effects in a particular class of data tends to end up owning that data simply through having the biggest pile, or the best results, not because they have unique data. (Again, Google: Microsoft and Yahoo! have the same data for the most part; Google is better at creating value for others from it.)

I will note that Dell recently announced that they were abandoning their long-heralded build-to-order methodology in favor of standardized commodity boxes. But I still stand by my statement.

There are two possibilities: first, that Dell is wrong, and their new supply chain approach will not save them, just make them more like everyone else. It could be that their “live suppy chain” approach just got too crufty, too complex – the article linked above suggests more than 5000 possible configurations. Maybe what they needed to do was to make the system smarter again by streamlining and simplifying.

But it’s possible too that the competitive advantage to be wrung from a live enterprise only takes you so far, and that in certain circumstances other advantages are more important. It may well be that the PC market has reduced itself to such commodity status that standardization trumps customization. It may well be that the costs of physical goods mean that the laws of virtual networks are only partly true in that realm.

I haven’t studied Dell’s situation and market sufficiently to have a fixed opinion. But the answer is knowable – a good field of study for business school cases. It’s worth repeating something I once said about open source, but now for Web 2.0: This is science, not ideology. Our goal is to understand what works, and why.

P.S. I also talked about the ideas here in my piece from last year, What Would Google Do?

tags: ,
• I wasn’t able to attend the Web 2.0 Expo so I am thrilled you wrote this follow up note. This past weekend I wrote a post for Fast Company that in a way builds on some of the ideas of enterprise 2.0 you shared here “How Micro-Targeting Can Help You Make Your Customers Happier” http://tinyurl.com/4nuc3y

Clearly there is more than one way of finding out what works. The important piece is looking and listening for it.

• Your three principles of Enterprise 2.0 also explain why today’s ad-based monetization approaches won’t work long-term for Web 2.0 properties. Brands graft their paid content poorly into sites, making commercial content stand out like a mismated arm on Frankenstein’s monster. Instead, companies should start internally and use existing data and process to uncover natural opportunities for Web 2.0 interaction, in line with business objectives.

• Perhaps it was as simple as Dell having looked at the data and found that most online customers bought standard configurations rather than customized them. With cheap components, the costs of customization may not make much sense compared to when components were expensive and trade-offs were worth making. If so, then I think you could argue that the collective knowledge of the ordering system showed that it could drop its complexity and eliminate itself, at least for retail customers.

The problem, as you point out, is that Dell will need to differentiate itself to keep viable. What might be interesting is to do something with Web 2.0 that helps the buyer/owner rather than Dell. For example, at a trivial level, just knowing which machines are most popular for which purposes, and how long they lasted before being replaced, might be of great help to a potential buyer. But what might really have potential is for Dell to build something that allows users to get more from their machines, perhaps through some networking software, that really provides value. Then they build value into the “Dell network” which might be the incentive to buy the brand, as opposed to a commodity machine with the same specs. A trivial example – suppose Dell installed WiFi at airports which allowed free access to Dell laptop owners? Wouldn’t that be of value to business travelers? Perhaps add services that could alert travelers to delays and advise on the successful plan changes of other travelers?

• Tim
The enormous amounts of data that is actually gathered from social interactions and every other user “contribution” as you mention results in an attempt to separate wheat from Chaff. Its a non trivial exercise. Just saying get better filters also does not solve this problem.

The reason why segmentation strategies work is because you can “bucket” sets of users into most likely to do “this”.

I believe the architecture, technology and systems do not exist to make user contribution a significant data point in strategy. Only trends can be included. Which is similar to what Wal Mart does. If they did to it with all the data points of user contribution, they would not be able to have sufficient time to process the information to make timely decisions.

• Sachin

@Cindy: You may be on to something when you say that Dell might have figured from usage data that people tend to go for standard configs. As Tim mentioned, it is a great subject for a case study to know if there is a point where the benefit is not worth the effort.

If this turns out to be true, there is a case to study also the long tail and whether it generates enough sales to make business sense to continue build-to-order in some fashion.

But I have a question for Tim: Wasn’t ERP supposed to bring many of the benefits that Web 2.0 as you define it brings now? I am thinking of Beehive and why Oracle thinks it has a good shot at being a force in this segment (i.e. building on its ERP experience).

• Tim,

I think another piece you may have missed was the opportunity for Enterprise2.0 to create low-cost access to highly targeted consumers (or even customers) for forward-looking market research. In other words, it’s not just the data being created by online consumers, it’s also the profiles that data creates that allows companies to deeply segment and then (in a low cost way) reach out and grab target consumers to validate and/or iterate on new initiatives, marketing campaigns, etc. Launching anything “new” is a huge gamble for companies and the ability of 2.0 to mitigate the risk is meaningful. Of course, there are more mundane interactions that are monetizable, too, as Yelp has shown by monetizing the dialogue between small business and customers.

• I’ve bought the 1521 laptop a couple of days ago. What can I say?
It’s perfect. The batteries works up to 4 hours (if you power off
the wireless interfaces). Also, I was lucky enough to get the model
without that TrueLife glossy screen. The bad thing is that the laptop
only comes with Windows Vista, and therefore has only Vista drivers.
It was somehow hard to find the Windows XP drivers for all presented
devices, but finally I have found them all.
Again – this laptop is just perfect.

• Yes I agree with Dell. You cannot find drivers for Windows XP, only for Vista and most of the people like to have Windows XP.

• Chinni Reddygari

Business Intelligence was supposed to take care of this analysis and provide recommendations. But for some reason it stalled – may be technology was not good enough but more than that I think people didn’t realize the potential and did not use all the inputs. BI 2.0 is trying to accomplish what BI didn’t and it’s clearly doable.