Danny makes a number of points in the article below. I’ll deal with them one by one.<?xml:namespace prefix = o ns = “urn:schemas-microsoft-com:office:office” />
So where does the smoking gun belong? I put it in the hands of RealNames itself. The company has made so many shifts in direction since it launched that it’s been hard to keep up with it.
Danny is correct that RealNames model changed several times. However – and this is a hard fact for many who have never run a business – business models are intended to change in the early stages of a startup. Every engagement with customers and partners teaches you much about the market. Shifts are inevitable.
Let’s examine the RealNames shifts and ask whether they were justified. Between April 1997 and March 1998 RealNames was busy building its Platform. Five employees [<?xml:namespace prefix = st1 ns = “urn:schemas-microsoft-com:office:smarttags” /><st1ersonName>Nico Popp</st1ersonName> and <st1ersonName>Bruce Ong</st1ersonName> from NeXt, Brian Russell – who traveled from the UK to join us, Myself and <st1ersonName>Gené McPherson</st1ersonName>] modeled the initial system.
Our goal was to become a layer of naming on top of the DNS. For this to work we needed to be available to users in the browser. We had discussions with Microsoft and Netscape. During 1998 we could not negotiate such deals and so we were forced to begin elsewhere. We built a plug-in and we made deals with AltaVista, LookSmart and Inktomi to include our results in their search engines. We sold Keywords to customers [generally non-generic terms so as not to conflict with the search engines results]. Initially we sold the Keywords for a flat $40 annual fee, but as our search distribution grew to include more than 50 search engines we moved to a pay per visitor, and often a pay per transaction model. This model survived through August 1999. As a company who had as its primary partners search engines it made sense to monetize search traffic.
By August 1999 two things happened. First, our vision of a layer of naming on top of the DNS became possible. After more than a year of discussions both Netscape and Microsoft agreed to embed RealNames in their respective browsers. This was a great breakthrough. Within days, AOL acquired Netscape and let us know it would not honor the agreement. But the Microsoft deal went through. At this time the deal only covered the USA and Microsoft retained the right to reject any Keyword it wished. Clearly this was not yet a great deal, but it was a beginning. Because we were now available on many more desktops we began charging increased amounts to the Fortune 2000 for Keyword packages. These deals were done through a direct sales force. We were able to deliver qualified visitors to our customers. Many deals in the $200k a year, through to a top deal of $700k a year were completed. The basis of sales was still Keywords, plus traffic. Revenue grew from $500k in 1998 to $3.6m in 1999, to $12.9m in 2000.
Between October 1999 and March 2000 we renegotiated our Microsoft contract and won support from Microsoft for a worldwide agreement for Keywords in the browser. Finally we had the Platform to be a naming layer on top of the DNS. Microsoft agreed to us selling Keywords under tightly controlled rules that, for example, excluded generic terms. But we believed that the global Platform for Keywords could now be built. This worldwide Platform went live in October 2000. It was such a big change in architecture it needed a change in the model. Rather than a small direct sales force selling to a handful of customers we needed a worldwide sales Platform selling to millions of customers.
In November 2000 we announced – at ICANN in LA – that we would be putting a global Keyword sales channel in place. It would be possible to buy Keywords from any one of hundreds of outlets. Between November 2000 and November 2001 this channel grew from none to over 1000 Keyword resellers. RealNames was not a Keyword monopoly but became the instigator of a Keyword market place. Typically 80% of the revenue from Keywords was enjoyed by the channel [ a country Registry and its Registrars]. VeriSign announced in October 2000 that its .com registry had made Keywords available for sale to ICANN accredited registrars.
Keywords were $50 per Keyword per year. There was no longer any payment for traffic. This was a mass market pricing plan similar to that in domain names. By March 2002 over 500 Keywords a day were being sold through this model. 2001 revenues were $13.9mm and 2002 was on track for a $24-28mm year.
In the summer of 2001 Microsoft was angered by the fact that many more Keywords were being sold and that some of them were of poor quality. This issue was addressed and an agreement was reached. There would be 2 products. Basic Keywords and Keywords Plus.
Basic Keywords would be sold on a first come-first served basis. This product would be available for all “low traffic terms” in MSN Search [less than 10 searches a month]. The price would be $50 a year. Keywords Plus would be $500 a year. This was for “high traffic terms”. Generics were not allowed. A human being would approve the ownership of these terms. A frame would be delivered with every Basic Keyword to allow the user to select search instead of the destination of the Keyword.
This change was forced on RealNames by Microsoft and we obligingly accepted it. We believed that it addressed all of Microsoft’s issues regarding quality.
So, in reality, not one of the changes was wrong. All were justified by the new Platform RealNames had built or because they were required by Microsoft in order for the Platform to exist. RealNames revenues grew every year. Annualized revenue per employee grew every month for 4 years too – reaching more than $200k per employee by March 2002. As Danny notes, we finally got to a sensible model, and one consistent with the vision. We also got cash flow positive. And VeriSign had agreed to Project Eureka, where more than 100 million Keywords would go live – 4 for all of the 25m domain names. It was at this point that – in my view – Microsoft’s fears of not controlling the technology led them to want to kill it.
So, change yes. Bad change. Not at all.
Danny’s second point is:
Unfortunately, RealNames also moved to a reseller model in 2001 that caused new confusion.
The reseller model was the only sensible model for a naming layer. It is reseller or monopoly. We chose reseller. I’m not sure this is worth arguing because it is clear that economies of scale come from maximizing scale. RealNames could not build a global sales infrastructure alone, even if we had wanted to be a monopoly. We needed help. Sure, some channel partners made claims to customers that were just plain wrong. It wasn’t many and it wasn’t often, and it was usually the same culprits. RealNames habitually cancelled reseller contracts if it became aware of such behavior. I believe we did a great job of democratizing the sale of Keywords. Microsoft would have preferred it to stay under the control of a single entity. Well – you choose who is right.
Finally Danny has this to say:
Clearly, it was the wrong gamble. But blaming Microsoft for this? Microsoft didn’t ask RealNames to become nearly 100 percent dependent upon it, any more than a casino asks you to bet all your money in a game of chance. The choice was one that RealNames made.
Well Danny, I’ll give you this one . Clearly it was a wrong gamble. Blaming Microsoft – absolutely! Sure they didn’t ask us to become dependent – although they did insist we discontinue the distribution of our plug-in client in August 1999. But IE’s very dominance means we – and others – have absolutely no choice, we are dependent. So we proclaimed to the world that it was possible to partner with Microsoft that they were capable of being fair minded. We gave them 20% of the company in return for their promises. We made them partners, not just business associates. We made contracts with third parties in good faith. Many of them built entire businesses on the back of that good faith. In one decision – driven by narrow product vision and a desire to control – Microsoft killed an entire market. That choice was made by Microsoft. It is a poor business decision made by soulless people who clearly have the arrogance that comes from absolute power.