Don Dodge on The Next Big Thing

Click fraud rates lower than some estimates

News.com reports " Click fraud rate lower than expected " based on a study by Click Forensics.

The Click Fraud Index shows that the overall, industry-wide average click fraud rate is 13.7 percent. The click fraud rate at top-tier search engines such as Google and Yahoo is even less, at 12.1 percent, the data show. The rate rises to 21.3 percent at so-called Tier 2 search providers and 29.8 percent at Tier 3 search companies, according to the Index.

Pundits including Henry Blodgett and Mark Cuban have long suggested that click fraud is rampant...approaching 30% to 50% of all clicks. Click fraud is when a web site owner clicks on ads on its own site to generate revenue for themselves, or when a company clicks on a competitors ads to drive up their costs. The claim is that they use automated software robots or pay people in third world countries to manually click on the ads.

The major search engines have sophisticated algorithms for detecting click fraud and remove it before it gets to their customers accounts. Based on my experience at AltaVista it is fairly easy to detect and disregard click fraud. You look for things like frequency of clicks per minute, number of clicks coming from the same IP address, seconds spent on the site, etc. This also explains why the Tier 1 sites have lower click fraud than the Tier 2 and 3 sites, The Tier 1 sites are more sophisticated and spend more money on click fraud detection and prevention.

Google recently settled a click fraud class action suit for $90M in advertising rebates. This settlement of $90M versus the billions of ad revenue indicates that click fraud is a much smaller problem than the alarmists have suggested. Good news for advertisers and search engines.

UPDATE 4/21/06 - Mark Cuban has entered the debate on click fraud. See his comment to this post below, and Mark's post on BlogMaverick. Mark makes good points and inspired me to add this update below. I also spoke by telephone to Mikhail from ClickFacts. They are still early in the process but do have some preliminary data that suggests click fraud varies greatly from site to site.

Spam stinks for sure...but it isn't click fraud. Spam and Splogs are designed to attract viewers with their "content" ...dubious as it may be. The sole purpose is get visitors to click on ads. I don't consider this to be click fraud since a real human is choosing to click on the ad.

Advertisers who pay for 'impressions' or page views, formerly known as CPM are definitely affected by spam and splogs. Spam pages designed to attract users by deception so that they can build of page view counts and then charge advertisers CPM rates are indeed a fraud. But that is not click fraud.

Blogs "essential" to a good career?

The Boston Globe has a story today Blogs "essential" to a good career with eight reasons why blogging can help your career, and several examples of people who have improved their lot in life through blogging.

Blogging, in my opinion, can be great for your career...or ruin you. It depends on how mature, articulate, and insightful you are. Teenagers who are "blogging" on MySpace today may later regret it when their potential employer does a web search for their name. Many adults would also be wise to think twice before they write.

Blogging is a natural, enjoyable, and positive experience for me. The feedback has been amazing, and has led to meeting new people, new learnings and deeper understanding. Writing forces me to think through my assumptions to a deeper level than I otherwise would, and improves my clarity of thought.

Blogging can certainly get you noticed...for good or bad. I have been contacted by lots of executive search people over the past 9 months including recruiters from Yahoo, Ask Jeeves, Disney, The Gap, and NBC. I have the best job in the world at Microsoft so the temptations have not swayed me.

The Boston Globe article lists 8 reasons why blogging is essential for your career. While I disagree that blogging is essential...if done well it can certainly help. Here are the 8 reasons;

1. Blogging creates a network.
2. Blogging can get you a job.
3. Blogging is great training.
4. Blogging helps you move up quickly.
5. Blogging makes self-employment easier.
6. Blogging provides more opportunities.
7. Blogging could be your big break.
8. Blogging makes the world a better place.

Tim Bray, who now works at Sun Microsystems, goes the Globe one or two better with 10 reasons why Blogging is Good.

1. You have to get noticed to get promoted.

2. You have to get noticed to get hired.

3. It really impresses people when you say “Oh, I’ve written about that, just google for XXX and I’m on the top page” or “Oh, just google my name.”

4. No matter how great you are, your career depends on communicating. The way to get better at anything, including communication, is by practicing. Blogging is good practice.

5. Bloggers are better-informed than non-bloggers. Knowing more is a career advantage.

6. Knowing more also means you’re more likely to hear about interesting jobs coming open.

7. Networking is good for your career. Blogging is a good way to meet people.

8. If you’re an engineer, blogging puts you in intimate contact with a worse-is-better 80/20 success story. Understanding this mode of technology adoption can only help you.

9. If you’re in marketing, you’ll need to understand how its rules are changing as a result of the current whirlwind, which nobody does, but bloggers are at least somewhat less baffled.

10. It’s a lot harder to fire someone who has a public voice, because it will be noticed.

Blogging is not for everyone. It takes a lot of time, thought, and practice to do it well. Blogging is good exercise for thinking critically and writing clearly. If you don't feel comfortable writing your own blog, start by commenting on other blogs. There are at least 35 million blogs out there so there is something for everyone. Enjoy!

On Demand software in demand?

BusinessWeek has a story today The On-Demand Software Scrum that is similar to my SaaS (Software as a Service) post. The BusinessWeek story focuses on Oracle[ORCL], Microsoft [MSFT], SAP [SAP] and others efforts to enter the SaaS / On-Demand market.

On-Demand or SaaS software is typically "rented" on a monthly subscription basis rather than purchased up front, and hosted and maintained by the software vendor, not the internal IT staff. According to AMR Research about 10% of software is sold this way now, and it is growing at about 20% per year.

Oracle expects that up to 50% of its software will be sold on a subscription basis going forward, much of that from recently acquired Siebel.

BusinessWeek points out that Siebel was a very successful CRM company that got squeezed on the high end by SAP, and attacked on the low end by Salesforce.com. Siebel sales slumped and after several years was acquired by Oracle.

SAP offers SaaS licensing for its products but believes it is an on-ramp to its higher priced traditional offerings. Microsoft is experimenting with SaaS through its Office Live and Windows Live services.

Oracle, Microsoft, IBM, SAP, and other traditional software companies are entering the SaaS / On-Demand market but face some strategic issues. Large public companies need to continue revenue growth to support stock prices. Sales people with big quotas will be reluctant to sell cheap monthly subscriptions.  Taking on the 24X7 hosting responsibility introduces new challenges and legal responsibilities.

Any company entering this market should consider the following issues. I originally wrote these for start-ups, but they also apply to large established companies.

Cashflow - VCs or investors must be patient and willing to fund you to break even.

Sales compensation - Look at metrics, total compensation, and cash flow.

Engineering - Have an open architecture that supports end user customization

Hosting - Partner with a solid, well funded, hosting company

Legal contracts - Automatic renewals, termination clauses, SLA (Service Level Agreements), privacy issues, data loss, data export, and a variety of other issues must be addressed in your subscription license.

Blogs double every 6 months, now 35M blogs

Dave Sifry, CEO of Technorati, published a report today State of the Blogosphere, April 2006. The number of blogs has been consistently doubling every 6 months for the past 3 1/2 years. Today there are 35M blogs compared to 16M just 6 months ago.

Given the law of large numbers, percentage growth will probably moderate over the next six months. In fact, it already has. I believe there were 27M blogs 3 months ago at the end of 2006, compared to 35M today. In order to continue the streak of doubling every 6 months there would need to be 54M blogs by the end of June. There were *only* 8 million blogs added last quarter so it is very unlikely there will be 19M more added in the second quarter of this year.

The law of large numbers applies to all fast growing companies as well. Google has experienced tremendous growth over the past 4 years, and will continue to grow in absolute terms. But, the percentage growth that Wall Street looks at will decline because the base denominator is getting so big. I wrote "Growing a Google Every Year...4 years in a row" which showed that Microsoft increased revenues by $4B per (more than Google's revenue for the whole year), yet it represented just 10% growth for Microsoft over the previous year. But I digress.

Bloggers are actively posting to these new blogs. About 20M people are still posting to their blogs 3 months after they were created. And about 4M post new articles at least once a week. Technorati estimates that 9% of new blogs are spam, and that 60% of "pings" are from known spam sources. Of course the most popular blogs get spammed the most causing some bloggers to disable comments and trackbacks. There is a business opportunity for an effective blog spam blocker.

Other interesting facts from the report;

Technorati now tracks over 35.3 Million blogs

The blogosphere is doubling in size every 6 months

It is now over 60 times bigger than it was 3 years ago

On average, a new weblog is created every second of every day

19.4 million bloggers (55%) are still posting 3 months after their blogs are created

Technorati tracks about 1.2 Million new blog posts each day, about 50,000 per hour

Social Search or Q&A Search?

BusinessWeek has a story today "Microsoft Gets Social" which talks about projects underway at Microsoft that incorporates social network search technology.

“Microsoft plans to unveil a question-and-answer social-search tool in the coming months, says Justin Osmer, senior product manager for MSN. The feature will let users direct questions to a specific universe, such as a group of friends, rather than to get automated lists of results from a generic search engine.”

Danny Sullivan of SearchEngineWatch points out "BusinessWeek confuses the forthcoming Q&A tool with social search. Those are two completely different types of search features/services, of which Q&A is the least important. In fact, it just reeks of another "me too" move that's not going to budge MSN Windows Live Search's usage among searchers."

I agree. Q&A Search involves asking a question of a group of users or "experts". I guess the group of users could be your social network...but I am unclear how this would work. Would your search be emailed to your social network group and wait for answers to come back? Who would be willing to wait for that? There have been several attempts at "Answers" or Q&A Search without much success.

Social Search typically means that the basic search results are filtered by the preferences or search history of your friends and social network. Jookster is a good example of this type of social search.

All of these approaches are attempts to improve relevance of search results. My experience is that search users are very impatient. They will not wait more than two seconds for results, or spend time organizing their search preferences, or enter a group of friends in a profile.

At AltaVista we had  search "word pairs" to help the user refine their search. If they entered a one word query like Jaguar, we would immediately serve up a results page for Jaguar, but also offer a line that said "Did you mean "Jaguar car", Jaguar animal", Jaguar NFL", etc. If a user entered a two word query we would offer alternative three word queries to help them further refine results.

There are lots of search companies trying to use tags, social networks, user history, voting schemes, and all sorts of things to try and improve search results. My guess is that users will get better at entering precise queries, and search engines will get better at ranking results, or offering alternative search queries. The search game is far from over.

Search engine users are quick to switch

The BBC reported on a  new survey by Jupiter Research which says search users will look at three pages of results (or less) before trying a new query or another search engine. Loyalty is fleeting. The survey found that 41% of users would change search engines, or search terms, if they didn't find what they were looking for on the first page. I suspect that most of that 41% simply try another search term before switching engines, but the report was not specific on that point.

The survey reported that 62% clicked on a result on the first page, and 90% clicked on a result within the first three pages.  Search users are getting better at entering more precise search terms, and search engines are getting better at delivering more relevant results. However, it is surprising to me how quickly users will switch search engines if they don't immediately find what they want. Google's current dominance is not assured.

When I was at AltaVista we ran statistics on all kinds of user behavior. I don't remember the exact numbers but a huge percentage, something like 75%, of all queries were two words or less. There are lots of ambiguous words or synonyms in the English language, so returning relevant results on a two word query was, and still is, extremely difficult.

Here are a few hints for effective searching.

• Use at least three descriptive and unique words in a query.
• Use "quotation marks" when you want only those words together. For example, the search query Don Dodge will find all instances of the letters don, including don't, donate, done, donor, etc., and all instances of the letters dodge. By placing "quotation marks" around "Don Dodge" you get only results that contain exactly those letters in that order.
• Use other words to limit or narrow your query. For example "Don Dodge" Microsoft, will return only results that contain "Don Dodge" and the word Microsoft somewhere in the result.
• Use the minus sign "-" to exclude certain results. The query "Don Dodge" - Microsoft, will return all instances of Don Dodge not related to Microsoft.

Most search engines have advanced search options that do a great job of helping you form more precise queries. Very few people use these features but they are very effective and don't take much extra time.

SaaS - new software model, new challenges

Software as a Service (SaaS), Software on Demand, Hosted Solutions...it goes by many names and is a new emerging software business model. The benefits to the buying customer are obvious...rent the software and hosting infrastructure rather than buy it up front. But, what happens to the software company business model? The answers are different for start-ups versus established software companies.

Start-ups need to get to revenue, and then to cash flow break even quickly...before the cash runs out. Any entrepreneur knows that cash is king. Selling up front perpetual licenses brings in the cash, but SaaS subscriptions provide a smooth predictable revenue stream that builds up over time.

Michael Skok of North Bridge Venture Partners says that it takes 70% to 100% more capital to fund a SaaS company to a liquidity event than a traditional perpetual license company. It also takes 2 to 3 times longer to get there. However, NBVP believes the extra time and money is worth it and pays off in higher market cap values.

Mr. Skok is not guessing...he has real numbers. NBVP has 8 funded SaaS companies out of a total of 53 portfolio companies. The estimates are averages for enterprise software applications. Consumer applications can be done less expensively, and get to market faster. The feature requirements, integration and customization requirements are also usually less demanding for consumer applications. NBVP actively invests in both enterprise and consumer SaaS companies.

Michael Skok was a panelist on a Mass Technology Leadership Council session on Software as a Service. He shared the following facts based on his investments in SaaS companies;

• SaaS companies need an average of $35M in VC capital, versus $20M for a similar perpetual license company.
• It takes 6 to 7 years to get to a liquidity event (IPO or acquisition)
• Public equity markets pay a 10% to 20% premium for predictable revenue streams
• SaaS companies move faster than big companies. They can introduce new features instantly versus waiting for the next major release. Think years.
• SaaS requires an architecture that supports end user customization
• Industry standards are critical for interoperability
• Steady state business models require 15-18% for engineering and 30-35% for Sales and Marketing.

Established traditional software companies must deal with "renting" their software for something like $5 per user/ per month versus selling it for $20K. This effects quarterly profitability and sales compensation. Established companies have the advantage of cash flow from their perpetual license business to sustain them while the SaaS business is growing, but the strategy conflicts are pervasive.

If you are a sales person with a $1M quota and you have the choice of selling a $100K perpetual license, or a $2K a month SaaS subscription, which would you do? Some companies are changing their sales compensation models and metrics to level the playing field. They might give quota credit for 24 to 36 months of the subscription and pay commissions as the revenue comes in. Or they might just pay commissions based on 18 to 24 months of subscriptions.

Engineering costs for SaaS software are higher for both start-ups and established companies, and it takes longer to develop. Engineers must design in user customization and remote development capabilities. It takes a unique architecture and extensive User Interface design to allow customers to customize and integrate a hosted SaaS application, especially remotely. Remember, this isn't shrink wrapped desktop software.  Most enterprise software is customized in some way for each customers environment. Traditional companies sell consulting services to do this. SaaS companies must build the customization features into the base software.

Hosting is a big deal, and a subject unto itself. I will not get into it here but suffice it to say that it is very expensive and complex to build and manage a 24X7X365 service that is always up. Most companies outsource this, but it is not free. It must be factored into your pricing model and cashflow model.

How do software companies price their SaaS offerings? Jim Geisman president of MarketShare, a software pricing consultancy, said most companies use a 3 year break-even formula for SaaS vs. Perpetual license prices. Jim makes his living consulting on these issues so I will not reveal all his secrets, but he did share this hypothetical example.

A SaaS company prices its product at $65 per user per month for 10 users. A competing company offers to sell you the same software for $17,500 plus the normal 15% annual maintenance. Which is the better deal? It depends on how many users and your time horizon. The break even point is 3 years. After that SaaS actually costs more than the traditional license. But, after three years you would probably upgrade your traditional license to a new release, so the amortization clock starts again.

The panel predicted that SaaS would continue to grow as a percentage of total software contracts, but that both models would persist for a long time. The estimates are that about 10% of software is sold as SaaS now and that over the next 5 to 10 years it will grow to 25%.

If you are considering the SaaS model for your software company make sure you focus on these things;

• Cashflow - VCs must be patient and willing to fund you to break even.
• Sales compensation - Look at metrics, total compensation, and cash flow.
• Engineering - Have an open architecture that supports end user customization
• Hosting - Partner with a solid, well funded, hosting company
• Legal contracts - Automatic renewals, termination clauses, SLA (Service Level Agreements), privacy issues, data loss, data export, and a variety of other issues must be addressed in your subscription license.

Done correctly SaaS is a very compelling and rewarding business model. Salesforce.com is the poster child for SaaS success. The stock market will reward companies that can get "over the hump" and build a predictable revenue stream. Smart VCs are willing to commit the funds necessary to get to liquidity.

The "Tom Sawyer's" of Open Source make millions while "contributors" work for free

Red Hat reportedly will acquire open source project JBoss for $350M. Ponderings of Woodrow included some interesting JBoss numbers from the Red Hat conference call.

CY2006 - In excess of $60 million bookings, $40 million revenues

CY2007 - $110 million bookings, $80 million revenues

150% YOY revenue growth in March Q

JBoss growing revenues faster than expenses

Cash flow negative historically (moving to positive in a few quarters)

$0.01 dilutive to Q2 (August) for RHAT

Neutral to earnings and cash flow for full fiscal year (Feb 07)

Dana Gardner who writes "Between The Lines" for ZDNet says this acquisition represents a seismic shift towards open source. Perhaps this "seismic shift" will have some unintended consequences.

How will the developer "contributors" who worked for free on the project react to a few people at the top making hundreds of millions? I am reminded of the famous Huckleberry Finn story where Tom Sawyer convinces his friends to work for free and paint the fence for him. "Come on it will be fun".

The Open Source phenomenon has been difficult for venture capitalists and business people to understand. Why would talented software developers work for free? When something of value is created who collects the rewards? How could open source be monetized?

Then there were the legal issues of ownership, patents, copyrights, etc. Enterprise IT managers worried about copyright or patent infringement lawsuits if they used open source project software with no warranties. Other concerns were who would fix bugs, add new features, or address performance, scalability, and integration issues?

Now that Open Source software has hit the big time and become a billion dollar business will things change? Will developers still be willing to "contribute" for free? Now that there are millions of dollars at stake will patent lawyers jump in with infringement lawsuits? When big companies buy open source project companies will the business model change? How will these big companies recover their investment?

IBM, HP, Oracle, Novell, and Red Hat were all playing together nicely in the open source world. Now that Red Hat and Oracle have made big acquisitions will they all continue to play nicely?

Did you ever think you would see The Adventures of Tom Sawyer applied to software?

RedHat acquires JBoss - open source consolidates

CNet reported today that Red Hat has acquired JBoss for approximately $350M. Open Source software is attracting lots of attention from the big players and VCs. There have been several acquisitions recently and probably a few more to come.

There was speculation that Oracle would acquire JBoss after recently acquiring open source database companies SleepyCat and InnoDB. The Open Source business is consolidating into a few big players. MySQL is probably the largest independent open source company remaining.

JBoss is an open source application server company founded by Marc Fleury. InfoWorld estimated that JBoss did about $20M in revenue last year and might do $50M this year. Assuming $50M revenues, and a $350M acquisition price, that yields a 7X forward revenues multiple. Marc Fleury reportedly owns 50% of the company. VCs David Skok of Matrix and Peter Fenton of Accel were lead investors in the company. JBoss has 150 employees.

The Open Source software market is consolidating. There will be a few winners and many "also-rans" that never really make it. In an earlier post "Open Source - the Darwinian approach to software" I talked about the odds of success for open source projects.

Open Source projects must pass the Darwinian test (only the fittest (most interesting) survive) to attract developers and contributors to the project. Maybe 1% of all Open Source projects attract more than 30 developer/contributors. Then the user community applies its own Darwinian test to that 1% of projects, as only the most useful and supported projects attracts a significant user community. The Open Source projects you read about are the ones that passed the Darwinian tests. Yet, the media would have you believe that Open Source is a sweeping movement.

There are VERY few VC investment opportunities in Open Source. There will only be one or two winners in each space, and the larger spaces are already covered. Linux has already been covered by Red Hat, IBM, HP, Novell, and others. Apache is also supported by many companies including Covalent and Microsoft. The database space is already covered by MySQL, Postgres, Ingres, SleepyCat, BerkleyDB, and 4dbObjects.

Oracle, Red Hat, IBM, Novell, and other large players are busy acquiring the most successful open source companies. How will this affect the open source community? Will the acquired companies remain open source? Will business models change? Would you bet your company on open source? Would you build mission critical applications on open source? Questions to consider for entrepreneurs and  IT executives.

27 business plans in one day

Earlier this week I attended the Mass Tech Leadership Annual Investment Conference and the Keiretsu Forum monthly screening meeting. During the course of the day I reviewed the business plans of 27 companies looking for funding. They were looking for $500K to $5M so most of them were looking for Angel money and not ready for VCs. The companies were all over the map in terms of products and services.

The list of companies is useful as a snap shot of where entrepreneurs are focusing their attention. In no particular order;

• Assembla - collaboration platform for enterprises
• Optodot - material sciences - nanoporous membranes
• bWary - configuration monitoring/compliance
• AnnouncementMail - email spam filter
• eRPortal - asset/material management
• xCellaSave - cell phone advertising
• InBoxer - email filtering and compliance
• Clinical Technology Advisors - patient compliance
• Beam Power - cell phone tower amplifiers
• MaestroTec - document management services
• RoamingWire - WiFi Hotspot hardware and software
• Jarg - semantic search technology
• AptSoft - transaction/event monitoring & notification
• Enventra - OCR for math equations
• RuleSphere - compliance rules engine
• MobileSecure - cell phone data security
• Kyos - digital health records - conversion/mgmt
• StarTech - Outsourcing software development
• InsightDirect - field service management software

InsightDirect was the best/most mature business I saw at the conference. They have a great product, solid management team, lots of customers, $3.6M in revenue last year, and a good business model.

RoamingWire could be in a great position too. They have very small WiFi antennas that mount on utility poles to create a WiFi Hotspot network.

MobileSecure is in a hot market - cell phone security. Could be big.

The Keiretsu Forum also had some interesting companies present. Here is a quick run down;

• Metaphor Solutions - IVR (Interactive Voice Response) for small businesses.
• Opthometrics - medical device to measure eye movements
• EggRock - Pre-assembled building components for hotels and universities
• Eastbridge Real Estate - Structured finance of mortgage securities
• Jookster - search engine which utilizes social networks for relevance
• Innovia Medical - device for detecting ear infections
• Cognistar - online training service
• Bezoloven - New drug delivery method for treating lead poisoning

Jookster is a search engine that refines results based on the  votes of your friends or social network.  They apply the normal ranking and relevance criteria and then filter the results based on your own history/preferences and those of your friends. This could take off if they can target the right market. Search is a hot market...anything can happen.

This day reminded me that there are still lots of problems to be solved. The key to success is the people, the product, and the market. There are always exceptions, but the probability of success goes way up if the people are experienced entrepreneurs, the product solves a real pain point, and the market is ready to accept the new solution. It sounds easy, but it is never obvious.