[ About Us | Popular | Marcom | AdNet | IChannel | Glossary ]

Aug 25, 2008

How to build a network

How to build a network

I was talking with a media exec who started a blog ad network — bless him — but who I thought was taking too high a share of the revenue: at least half. That’s a natural reflex, perfectly understandable: Get what you can. Other networks do that. But the problem for me is that by taking too much, he excluded me from the network — I’m sticking with BlogAds, which takes only 20 percent — and the problem for him, then, is that this slows the growth of his network and a smaller network is a less valuable network. He understands this will because he’s a smart media guy. He wants a large network.

I was actually just trying to channel the network wisdom of Yochai Benkler, author of The Wealth of Networks, and Tom Evslin, who had talked about how to grow networks at a Union Square roundtable about collaborative production more than a year ago. I wasn’t sure I was getting it right, so I went to Evslin’s blog and asked him for a reprise, which he has just provided, brilliantly. I’ll summarize:

The first counterintuitive lesson: Companies that build large networks on the web don’t charge users what the market would bear; they charge as little as they could bear. That is how they maximize growth and value for everyone in the network on top of the platforms they provide.

In his blog post, Evslin takes this a step farther, pointing out that if you run a network that depends on scale, such as an ad network, then the more pages you have to sell, the bigger and better advertisers you can attract and the more you can charge. So if you take a smaller commission for each ad in the network, more sites will join it with more pages, which can now be sold at a higher value.

It gets even more head-scratching: Evslin argues that if you are too profitable, then you will attract competitors who will undercut you and steal market share. “If you’re doing well but running at or close to breakeven,” he explained, “you’ve made it impossible for anybody to undercut you without running at a deficit, which is hard to get funding for.”

So, to sum up: Take the minimum value out of the network to make it grow to maximum size to enable its members to charge more for their value while keeping costs and margins low to block competitors.

That’s not how old networks operate. Cable companies wrap their wires around you and squeeze maximum fees out of you. Ditto phone companies, newspapers, and retailers. But they all face competition from next-generation networks.

craigslist is the Evslin poster child. It foregoes revenue for most listings in most markets—charging just for job ads and for real estate in a few markets—and that turned it into the critical-mass marketplace for most listings. “If Craig now attempted to maximize revenue by charging for a substantially higher percentage of ads, a door would be cracked open for competition,” Evslin writes. “There is no chance at current rates for a competitor to steal Craig’s listings (and readers) by charging less.”

I’m writing about the network model in the book and this will also be a key topic of discussion in our event at CUNY on new business models for news; that’s why I’m talking about it.

15 Responses to “How to build a network”

  1. Julie Kentwood Says: 
  2. At what point do you switch over and start incrementally extracting value from the network?

  3. Andrew Tyndall Says: 

    Funnily enough, old media networks seem to follow this plan — maximizing size while minimizing margins. It is my understanding that the broadcast television networks come close to merely breaking even in the business of selling advertising for revenue and at the cost of providing programing for their affiliates. Their business plan is that they happen to own and operate television stations in the largest and most profitable local markets. Thus their model makes money for all affiliates yet disproportionately benefits the affiliates the network itself happens to own.

  4. slight.frost.byte » Blog Archive » Optimize for Growth Rather than Revenue Says: 

    [...] up-front (little revenue), the overall potential for revenue is greater due to a larger audience. Jeff Jarvis does a great job summarizing what Tom Evslin has to say about this concept, which definitely holds [...]

  5. Eric Berlin Says: 

    It seems to me that your argument holds if you consider all inventory to be equal i.e. an impression is an impression is…

    But if you bring in the factor that some inventory could be more valuable than others, and then conclude that it’s possible to create value around that higher value inventory for both publishers (higher CPMs = more revenue) and advertisers (higher value ad placement = money well spent), doesn’t the exact opposite of your argument start to play out?

  6. Jeff Jarvis Says: 

    I don’t think so. Tom has that built in.
    There are two pricing variables: the CPM of the ad and the commission the network takes. The higher the CPM, the better for everyone; Tom argues that the larger the network, the higher the CPM can be because bigger, better advertisers can be approached.
    The commission is a separate line. He argues that minimizing that to maximize inventory acts to increase CPM. The separate consideration is his point about profit as a barrier. If the network takes too much commission then another network can easily start and undercut it; if it minimizes commission, then it’s not worth it for a competitor to enter.

  7. Eric Berlin Says: 

    I see Jeff, thanks. I think that in reality there are more than those two variables, but I appreciate the explanation.

  8. Eric Nehrlich Says: 

    I liked Hal Varian’s take on this in Information Rules (my review is here). Varian is a Berkeley professor of economics who’s now Google’s Chief Economist, and in the book, he makes exactly this observation that the value is in growing the network, so companies need to avoid the temptation towards greedily extracting value. He also makes the same point that trying to extract that value opens up the potential for competition. The amazing thing to me is that the book was published in 1999, well before most people were thinking about these sorts of issues.

  9. Andy Freeman Says: 

    > The amazing thing to me is that the book was published in 1999, well before most people were thinking about these sorts of issues.

    Pricing to discourage potential competition and network effects are old economic principles. However, saying that standard economics applied to the web was an amazing thing to say in 1999.

  10. Printed Matters » Connecting the dots for AP 2.0 Says: 

    [...] with all news providers, big and small. Because we don’t need another walled garden. And the value of a network increases exponentially with size. I think there’s a law about [...]

  11. henry Says: 

    Great post Jeff.

    I don’t get it, since we run a nicely profitable business that has sent bloggers 10s of millions of dollars.

    I think that other folks are greedy.

    We’re happy at 20% (slightly higher for newer bloggers) because we’re located in podunk Carrboro, NC (much lower cost of living than major metro areas), and we focus on building/maintaining relationships online, rather than employing high-churn, high-salary, low-satisfaction staff in NYC, LA, SF and Chicago.

    Incredibly, even Google seems to be taking close to 50% on some of its stuff. I guess they’re a lot less automated than they look.

  12. Serge Lescouarnec Says: 


    I like Henry’s take on the online state of affairs yet I was never able to get ‘BlogAds’ going on ‘Serge the Concierge’… (http://www.sergetheconcierge.com)
    OK I am a small fish in a big pond but I do not feel like ‘BlogAds’ always follows up on things.

    As for networks and all, I recently decided to give a go to Twitter and when checking your page realized that while your followers (not a sect) numbered 1400 plus, you followed only 2 people, yourself and your son Jake.

    I know life is busy and all but isn’t it more than a bit lopsided?

    My 5 centimes

    ‘The French Guy from New Jersey’

  13. Preston Says: 

    You raise an interesting point, but there’s a flip side to this discussion. I remember a professor once told me: “There can be only one Wal-Mart. Only one.” There can be only one low-cost provider. Everyone else needs to, if they’re going to operate in a similar space, differentiate.

    A competitor may not be able to steal Craig’s listings on cost, but someone could steal Craig’s listings if they provide (differentiate with) something that customers are interested in.

  14. Steve Hochschild Says: 

    We must remember that Tom Evslin was the original email guy at Microsoft, and he clearly learned all of this while there.

    Some of us remember when Tom and Microsoft destroyed the email business when they bundled their email application in with Windows for Workgroups, and within just a couple of years, the number of email user agent vendors went from hundreds to fewer than 5.

    His approach is fine when your goal is to be the monopoly player — make a business at such a low margin that you drive out all the others, then reap monopoly profits. Not that this is necessarily his goal…

  15. Dash Chang Says: 

    I’ve known Tom since the early 80’s. Although I admire his work, this commune-istic view of networks seems naive.

    1. His loss leader approach to starting a network is one strategy; but needs to be sustained by a service where customers pay market prices. Google offers a lot of services free, but AdWords pay the bills. Further, Google optimizes to increase CPC.

    2. Size for size sake works in certain industries, but not advertising nor communities. Publishing has always been about specialized niches that sustain high CPMs. Yahoo has the problem that they are one undifferentiated brand. How many advertisers want that global reach? Not many. Thus supply exceeds demand.

    3. Sell at cost plus to minimize competiors. Presumably, if you are the low cost operator, like Taiwan, you win. Microsoft and Google create complexity and patents to block competitors. Drug industries hide behind regulations to protect profits. Innovation protected by intellectual property makes more sense.

    A network needs to bind birds-of-a-feather where community members have economic or emotional reasons to stay together. A network platform can enable thousands of communities. Low price, low margins, and growth for growth sake does not apply to network growth.

    Just my two-cents.

No comments:

Post a Comment

Comments accepted immediately, but moderated.

Support Our Sponsors: