If you can’t see the video click here
Subscribe for more videos than you can ever watch here
What makes startups succeed or fail? More than 90% of startups fail, due primarily to self-destruction rather than competition. For the less than 10% of startups that do succeed, most encounter several near death experiences along the way. Simply put, while we now have some good theory, we just are not very good at creating startups yet. After 50 years of technology entrepreneurship it’s still an art.
Three months ago I wrote about my ex-student Max Marmer and the Startup Genome Project. They’ve been attempting to quantify the art. They believe that they can crack the code of innovation and turn entrepreneurship into a science if they had hard data rather than speculation of why startups succeed or fail. Max and his partners had interviewed and analyzed over 650 early-stage Internet startups. In May they released the first Startup Genome Report— an in-depth analysis on what makes early-stage Internet startups successful.
Now 90 days later Max and his team have gathered data on 3200 startups and they believe they’ve discovered the most common reason startups fail.
Today you’re invited to benchmark your own internet startup and see how you compare to the winners.
Benchmarking Your Startup
I hadn’t heard from Max for awhile so I thought he took the summer off. I should have known better, it turned out he was hard at work.
Max and his team built a website called the Startup Genome Compass (their benchmarking web site) that allows an Internet startup to evaluate their business performance. The Startup Genome Compass uses a hybrid “Stage and Type” model that describes how startups progress through their business development lifecycle.
The benchmark takes 20 or so minutes to go through as series of questions, and in the end it spits out an analysis of how you are doing.
The benchmark is not perfect, it may even be flawed, but it is head and shoulders above what we have now – which is nothing – for giving Internet startups founders specific advice on best practices. If you have a few world-class VC’s on your board you’re probably getting this advice in person. If you’re like thousands of other startups struggling to get started, it’s worth a look.
It’s Not How Big It Is – It’s How Well It Performs
If you’re interested (and you should be) in how you compare to other early stage ventures, they summarized their results in a report “Startup Genome Report Extra: Premature Scaling.”
One of the biggest surprises is that success isn’t about size – of team or funding. It turns out Premature Scaling is the leading cause of hemorrhaging cash in a startup – and death. In fact:
The last time I wrote about Max I said, “I can’t wait to see what Max does by the time he’s 21.” Turns out his birthday is in a week, September 7th.
Happy birthday Max.
Listen to the post here: Download the Podcast here
Describing your product as “new and “never been done before” instead of “we’re just like those others guys, but better” could cost your company billions. RIM and TiVo are two examples of getting it right and wrong.
Research in Motion (RIM)
By 1992 Research in Motion (RIM) had been in business for eight years, had 16 employees, sales of about $500,000 a year, and three or four business lines. That year the two founders decided to get serious about being a company, and hired a CEO. Soon, RIM was focusing on making products for people on the move, using wireless communication and digital data.
In the early 1990’s two different trends were occurring in wireless communication. First, wireless voice networks – cell phone networks – had started to emerge. The ability to make a phone call untethered from a traditional phone was revolutionary and was starting to catch on fast. These new cellular phone networks were built around two-way circuit switched technology designed to move voice calls without interruption.
At the same time, digital data networks to support “pagers” were also growing rapidly. Pagers were small receive-only devices with 1 or 2-line displays that showed the phone number of who was “paging” them. Users ran to a traditional telephone and called a paging service who would read them their message. Doctors and drug dealers equally found these devices handy. Unlike the circuit-switched cell phone networks, pager networks were built around digital packet-switched technology.
Sell Directly to Businesses
In 1996 RIM was still in the hardware business selling packet-switched wireless radio modems to OEMs. In a major strategy shift, they decided to sell a product directly to businesses. In 1997, RIM introduced the first packet-switched messaging device. It used narrowband PCS and was housed in a clamshell device with a full keyboard.
The new device could hold names, email addresses, phone and fax numbers and incoming and outgoing messages. In 1998 RIM quickly followed this up with a next generation product with an 8-line display, ran on AA batteries and would last 500 hours.
The fact that you could send messages interactively blew people away. Underneath the hood RIM’s product was a technical tour de force. But RIM decided to hide all of that from their customers.
RIM positioned the Blackberry as an “interactive pager” because pagers were something people could understand. While the device was actually was doing email, people understood it as “the pager that you could respond with.” While phrases like “mobile email and packet switching” didn’t mean a thing to RIM’s first customers, the “interactive pager” positioning proved important in attracting early adopters.
Resegmenting an Existing Market
RIM’s product needed very little explanation. If you knew what a pager was, you knew what an interactive pager was. You got it. (You might gulp at the price – paging prices were dropping like a stone ($9/month versus $99/month for a RIM interactive pager) since most people were moving from pagers to cell phone to get calls. But to businesses where instant information gave you a critical edge (Wall Street, politicians, etc.) these new capabilities were worth almost any price.
In today’s language of Customer Development, RIM positioned the Blackberry as a segment of an existing market – pager users who needed two-way communication. Their intent: initial sales would come from users who already understood what the product could do so adoption would occur rapidly.
RIM, the Blackberry and its network had more inventions per square inch than most startups. The founders could have easily described the product as “the first packet-switched interactive messaging network.” Or they could have said, “corporate email now seamlessly forwarded from your company’s network to your pocket.” They did none of that. The founders swallowed their pride and simply introduced the Blackberry as an “interactive pager.” Their board, with no need to prove how smart and creative they were, agreed.
After a few years, as users became comfortable with the technology, the entire space of interactive pagers became known as the “Blackberry or “wireless email” market rather than the “interactive pager” market.
In 1999, about the same time RIM introduced its first interactive pager, another advanced technology company, TiVo, shipped its first product.
Recording video on magnetic tape was developed in the mid 1950’s by Ampex, and had evolved into a consumer-friendly cassette by the late 1960’s. VCR’s caught on in the home in the late 1970’s driven by movie rentals and pornography. Sales of VHS-based VCRs exploded after Sony and JVC fought a brutal standards battle (Betamax versus VHS) and when the U.S. Supreme Court ruled that home taping of television programs for later viewing (“time-shifting”) constituted a fair use.
But cassette tapes were still bulky and awkward. And most consumers had never mastered recording a TV program (let alone setting the clock on their VCR.)
TiVo solved all those problems. It was the logical marriage of computers and video recording. Essentially TiVo was a computer with a hard drive integrated with a TV tuner and MPEG decoder. It digitized and compressed analog video from an antenna, cable or direct broadcast satellite. But it was the software that made the TiVo great. It was reliable. Its user interface was simple. It let users record from the familiar program guide. Since you were recording video to a hard disk, you could appear to pause live TV, instant replay, rewind or record anything.
TiVo originally sold directly to consumers through consumer electronics stores, via Sony and Phillips and was integrated into set-top boxes from DirecTV.
Creating a New Market
TiVo’s product needed very little explanation. After a demo, if you knew what a VCR was you knew what a TiVo was. You got it. (You might pause at the price – VCR prices were plummeting – $150 versus $800 for the first TiVos, but compared to a VCR it took your breath away.)
Except there was one problem. The TiVo CEO hated the idea that customers might think of TiVo as a better VCR. In fact he said, “Anytime anyone says that to me, I go completely nuts. So we had this challenge of explaining, It’s actually not a VCR. It’s a lot more sophisticated and uses a hard disk, and therefore you can record and playback simultaneously and do clever things like pause live TV, and so on.” And the board, being enamored with Silicon Valley technology, first mover advantage and concerned about the huge price gap between a VCR and TiVo, agreed.
As a result, the company instead chose to position TiVo as a New Market. In a new market when customers have no idea what the product can do, a company needs to educate potential customers about the space not the product. This results in a much slower adoption curve – the classic hockey stick.
TiVo spent the next five years trying to convince users that the box they wanted to buy as a better VCR was really something different. Hundreds of millions of dollars went into marketing campaigns to create an entirely new consumer electronics category – Digital Video Recorders. TiVo was first positioned as a “personal television system.” But no one knew what that meant. Next they tried the slogan “TiVo, TV your way.” Early adopters simply ignored the company’s positioning buying the device in spite of the inane descriptions.
But trying to create a totally new market took its toll. TiVo had plenty of other battles to fight: competition, issues with channel partners, patent battles, as well as the movie studios, cable companies, broadcast networks and advertisers who all wanted TiVo dead. Instead the company used most its cash on marketing and advertising in trying to define a new product category and accelerate adoption.
RIM sales were $15 billion in 2010. In the last ten years they’ve made over $9 billion in profit.
TiVo sales were $240 million in 2010. In the last ten years they lost $400 million dollars.
How much of this can be traced back to the time, money and energy they spent on their initial positioning?
- Market Type matters
- No one will stop you from picking a new market.
- If you do, realize you have defined a space with no customers. You now need to spend your marketing dollars in educating users about the market not your product.
- In an existing market you’ve picked a space that has customers. Here you need to spend your marketing dollars differentiating your product from the incumbents. Are you faster and better? Are you cheaper? Do you uniquely appeal to a segment?
First-Mover Advantage is an idea that just won’t die. I hear it from every class of students, and each time I try to put a stake through its heart.
Here’s one more attempt in trying to explain why confusing testosterone with strategy is a bad idea.
First mover advantage – great bad idea
The phrase “first mover advantage” was first popularized in a 1988 paper by a Stanford Business School professor, David Montgomery, and his co-author, Marvin Lieberman.
This one phrase became the theoretical underpinning of the out-of-control spending of startups during the dot-com bubble. Over time the idea that winners in new markets are the ones who have been the first (not just early) entrants into their categories became unchallenged conventional wisdom in Silicon Valley. The only problem is that it’s simply not true.
The irony is that in a retrospective paper ten years later (1998),  the authors backed off from their claims. By then it was too late. Using this idea to differentiate themselves as the hot new Silicon Valley VCs, some of his former business school students made this phrase their rallying cry. Soon every other VC was using the phrase to justify the reckless “get big fast” strategies of dot-com startups during the Internet Bubble.
Fast Followers – a better idea
In fact, a 1993 paper by Peter N. Golder and Gerard J. Tellis had a much more accurate description of what happens to startup companies entering new markets. In their analysis Golder and Tellis found almost half of the market pioneers (First Movers) in their sample of 500 brands in 50 product categories failed. Even worse, the survivors’ mean market share was lower than found in other studies. Further, their study shows early market leaders (Fast Followers) have much greater long-term success; those in their sample entered the market an average of thirteen years later than the pioneers. What’s directly relevant from their work is a hierarchy showing what being first actually means for startups entering new or resegmented markets:
|Innovator||First to develop or patent an idea|
|Product Pioneer||First to have a working model|
|First Mover||First to sell the product||47% failure rate|
|Fast Follower||Entered early but not first||8% failure rate|
The Race to Fail First
What this means is that first mover advantage (in the sense of literally trying to be the first one on a shelf or with a press release) is not real, and the race to be the first company into a new market can be destructive. Therefore, startups whose mantra is “we have to be first to market” usually lose. What startups lose sight of is there are very few cases where a second, third, or even tenth entrant cannot become a profitable or even dominant player. (The rules are different in the life-sciences arena.)
Ford vs. GM, Overture vs. Google
For example, Ford was the first successfully mass produced car in the United States. In 1921, Ford sold 900,000 Model Ts for 60 percent market share compared to General Motors 61,000 Chevys, a 6 percent market share. Over the next ten years, while Ford focused on cost reductions, General Motors built a diverse and differentiated product line. By 1931 GM had 31% of the market to Ford’s 28%, a lead it has never relinquished. Just to make the point that markets are never static, Toyota, a company that sold its first car designed for the US market in 1964, is poised to surpass GM as the leader in the US market. The issue is not being first to market, but understanding the type of market your company is going to enter.
If the car business is too removed from high tech as an example, how about the story of Overture. In 1998 Goto.com, a small startup (later Overture, now part of Yahoo!), created the pay per click search engine and advertising system and demo’d it at the TED conference.
It was not until October 2000 that Google offered its version of a pay per click advertising system -AdWords -allowing advertisers to create text ads for placement on the Google search engine.
Google is a $25 billion dollar company with most of its revenue from AdWords.
Overture was acquired by Yahoo for $1.6 billion.
Implicit Customer Discovery and Validation in Fast Followers
Why do fast followers win more often? It’s pretty simple. First Movers tend to launch without really fully understanding customer problems or the product features that solve those problems. They guess at their business model and then do premature, loud and aggressive Public Relations hype and early company launches and quickly burn through their cash.. This is a great strategy if there’s a bubble occuring in your market or you are going to bet it all on flipping your company for a sale. Otherwise the jury is in. There’s no advantage. 
Astute fast-followers recognize that part of Customer Discovery is learning from the first-mover by looking at the arrows in their backs. Then avoiding them.
- Believing in First Mover Advantage implies you understand your business model, customers problems and the features needed to solve those problems.
- That’s unlikely.
- Therefore you’re either going to burn through your cash or pray that the hype can help you can flip your company.
- None of the market leaders in technology were the first movers
 P. N. Golder and G. J. Tellis. 1993. “Pioneer Advantage: Marketing Logic or Marketing Legend?” Journal of Marketing Research, 30(2):158–170.
 Did First-Mover Advantage Survive the Dot-Com Crash? . M. Lieberman 2007
Listen to the post here: Download the Podcasts here
Trading emails with a startup CEO building an iPhone app, I asked him why potential customers would buy his product. In response he sent me a competitive analysis. It looked like every competitive analysis I had done for 20 years, (ok maybe better.)
And it made me sad. Looking at the spreadsheet, I realized that competitive analysis tables are one of the ways professional marketers screw up startups from day one. And I had done my share.
Prove What I Already Believe
Most competitive analyses are: 1) sales documents for investors and/or 2) an attempt to rationalize the founders assumptions.
It’s Part of the Plan
Most investors require you to write a business plan which includes a section called a “competitive analysis” in which you tell potential investors how your product compares to products other companies trying to develop and sell to the same customers. While most investors don’t actually read your business plan for a first meeting, a summary of your competitive analysis usually ends up as a slide or two in your PowerPoint presentation.
Your goal in this slide is to tell investors: 1) you understand the market you are selling to, 2) you understand the other companies selling in your market, and 3) you understand how and why you are better than any of the products currently in the market. You are also implicitly telling potential investors, “These features on our competitive slide mean we will sell a lot of what we are planning to build so invest in us.”
Death by Analysis
I looked at the competitive analysis this startup CEO sent to me. This guy was experienced, he worked at lots of large companies, so the table was thorough, it had lots of rows and mentioned all the competitors.
Not only was it wrong, it would set his company back months and possibly even kill them.
Competitive Analysis Drives Feature Sprawl
In most startups the competitive analysis feature comparison ends up morphing into the Marketing Requirements Document that gets handed to engineering. The mandate becomes; “Our competitors have these features so our startup needs them too. Get to work and add all of these for first customer ship.”
Product development salutes and gets to work building the product. Only after the product ships does the company find out that customers couldn’t have cared less about most of the bells and whistles.
Instead of optimizing for a minimum feature set (that had been defined by customers) a competitive analysis drives a maximum feature set.
This is not good.
Where Are the Customers?
Here’s the problem: How did the founder know which features to choose on the competitive analysis table? When I was running marketing, the answer usually was, “We’ll put up whatever axes or feature comparisons that make us look best in this segment to potential investors. What else would you choose?”
At its best a competitive analysis assumes that you know why customers are going to buy your product. At its worst it exists to rationalize the founder’s assumptions about what they are building. This is a mistake – and it is a contributing factor (if not a root cause) of why most startups get their initial feature set wrong.
If you are building a competitive analysis table, do so only after you understand that the features you are listing matter to customers. Most marketers are happy to build feature comparisons. But customers don’t buy features, they usually buy something that solves a real or perceived need. That’s the comparison you and your investors should be looking at – what do customers say they need or want?
The answer to that question is almost never in your building.
How to Make A Competitive Analysis Useful
A competitive analysis makes sense when your startup is entering an Existing market – where the competitors are known, the customers are known, and most importantly – the basis of competition is known.
(The basis of competition are the features that customers in an existing market have said, “Yes, this is what is extremely important to me. I will dump my current supplier/manufacturer for your new product because yours is smaller/faster/easier to buy/get to/tastes better, etc.)
You win in an existing market when you are better or faster on those metrics that customers have told you are the basis of competition. Your competitive analysis must be around those metrics.
But most startups are not entering an Existing market. They may be trying to:
In a Resegmented market, a competitive analysis starts with the hypothesis of “Here’s the problem we are solving for customers.” The competitive analysis chart highlights the product features that differentiate your startup from the existing market incumbents because of your understanding of specific customer needs (not your opinion) in this niche.
In a New market a competitive analysis starts with the hypothesis of “We are creating something that never existed before for customers.” The competitive analysis table highlights the product features that show what customers could never do before. It compares your company to groups of products or services.
I asked the CEO to go back to the competitive analysis and tell me whether he really knew what features matter most to potential customers. If not, he should get out of the building and find out.
- Too often competitive analysis drives product requirements in startups.
- This can lead engineering to build the maximum feature set rather than minimum feature set.
- You need to get outside the building and figure out what features matter to most customers.
- No feature lists without facts.
In my last post I described what happened when a company prematurely scales sales and marketing before adequately testing its hypotheses in Customer Discovery. You would think that would be enough to get wrong, but entrepreneurs and investors compound this problem by assuming that all startups grow and scale by executing the Revenue Plan.
The Appendix of your business plan has one of the leading cause of death of startups: the financial spreadsheets you attached as your Income Statement, Balance Sheets and Cash Flow Statements.
Reality Meets the Plan
I got to see this first hand as an observer at a board meeting I wish I could have skipped.
We were at the board meeting of company building a radically new type of communication hardware. The company was going through some tough times. It had taken the company almost twice as long as planned to get their product out the door. But that wasn’t what the heat being generated at this board meeting was about. All discussion focused on “missing the revenue plan.”
Spread out in front of everyone around the conference table were the latest Income Statement, Balance Sheets and Cash Flow Statements. The VC’s were very concerned that the revenue the financial plan called for wasn’t being delivered by the sales team. They were also looking at the Cash Flow Statement and expressed their concern (i.e. raised their voices in a annoyed investor tone) that the headcount and its attendant burn rate combined with the lack of revenue meant the company would run out of money much sooner than anyone planned.
Lets Try to Make the World Match Our Spreadsheet
The VC’s concluded that the company needed to change direction and act aggressively to increase revenue so the company could “make the plan.” They told the CEO (who was the technical founder) that the sales team should focus on “other markets.” Another VC added that engineering should redesign the product to meet the price and performance of current users in an adjacent market.
The founder was doing his best to try to explain that his vision today was the same as when he pitched the company to the VC’s and when they funded the company. He said, “I told you it was going to take it least five years for the underlying industry infrastructure to mature, and that we had to convince OEMs to design in our product. All this takes time.” But the VC’s kept coming back to the lack of adoption of the product, the floundering sales force, the burn rate – and “the plan.”
Given the tongue-lashing the VC’s were giving the CEO and the VP of Sales, you would have thought that selling the product was something any high-school kid could have done.
What went wrong?
Revenue Plan Needs to Match Market Type
What went wrong was that the founder had built a product for a New Market and the VC’s allowed him to execute, hire and burn cash like he was in an Existing Market.
The failure of this company’s strategy happened almost the day the company was funded.
Make the VC’s Happy – Tell Them It’s a Big Market
There’s a common refrain that VC’s want to invest in large markets >$500Million and see companies that can generate $100M/year in revenue by year five. Enough entrepreneurs have heard this mantra that they put together their revenue plan working backwards from this goal. This may actually work if you’re in an existing market where customers understand what the product does and how to compare it with products that currently exist. The company I observed had in fact hired a VP of Sales from a competitor and staffed their sales and marketing team with people from an existing market.
The VC’s had assumed that the revenue plan for this new product would look like a straight linear growth line. They expected that sales should be growing incrementally each month and quarter.
Why did the VC’s make this assumption? Because the company’s initial revenue plan (the spreadsheet the founders attached to the business plan) said so.
What Market Type Are We?
Had the company been in an Existing Market, this would have been a reasonable expectation.
But no one (founders, management, investors) bothered to really dig deep into whether that sales and marketing strategy matched the technical founder’s vision or implementation. Because that’s not what the founders had built. They had designed something much, much better – and much worse.
The New Market
The founders had actually built a new class of communication hardware, something the industry had never seen before. It was going to be the right product – someday – but right now it was not the mainstream.
This meant that their revenue plan had been a fantasy from day one. There was no chance their revenue was going to grow like the nice straight line of an existing market. More than likely the revenue projection would resemble the hockey stick like the graph on the right.
(The small hump in year 1 is from the early adopters who buy one of anything. The flat part of the graph, years 1 to 4 is the Death Valley many companies never leave.)
Companies in New Markets who hire and execute like they’re in an Existing Market burn through their cash and go out of business.
Inexperienced Founders and Investors
I realized I was watching the consequences of Catch 22 of fundraising. Most experienced investors would have understood new markets take time, money and patience. This board had relatively young partners who hadn’t quite grasped the consequences of what they had funded and had allowed the founder to execute a revenue plan that couldn’t be met.
Six months later the VC’s were still at the board table but the founder was not.
- Customers don’t read your revenue plan.
- Market Type matters. It affects timing of revenue, timing of spending to create demand, etc.
- Make sure your revenue and spending plan matches your Market Type.
- Make sure the founders and VC’s agree on Market Type strategy.
One of the benefits of teaching is that it forces me to get smarter. I was in New York last week with my class at Columbia University and several events made me realize that the Customer Development model needs to better describe its fit with web-based businesses.
Dancing Around the Question
Union Square Ventures was kind enough to sponsor a meetup the night before my class. In it, I got asked a question I often hear: “What if we have a web-based business that doesn’t have revenue or paying customers? What metrics do we use to see if we learned enough in Customer Discovery? And without revenue how do we know if we achieved product/market fit to exit Customer Validation?”
I gave my boilerplate answer, “I’m a product guy and I tend to invest and look at deals that have measurable revenue metrics. However the Customer Development Model and the Lean Startup work equally well for startups on the web. Dave McClure has some great metrics…” It was an honest but vaguely unsatisfying answer.
Union Square Ventures
The next morning I got to spend time with Brad Burnham, partner at Union Square Ventures talking about their investment strategy and insights about web-based businesses. Bill and his partner Fred Wilson have invested in ~30 or so companies with 27 still active.
They’re putting money into web services/business – most without early revenue. It’s an impressive portfolio. By the time the meeting was over I left wondering whether the Customer Development model would help or hinder their companies.
Eric Ries in Times Square
For any model to be useful it has to predict what happens in the real world – including the web. I realized the Customer Development model needs to be clearer in what exactly a startup is supposed to do, regardless of the business model.
What we concluded is that the Customer Development model needs an additional overlay.
Just as a reminder, the Customer Development has four simple steps: Discovery, Validation, Creation and Company Building. But it also requires you to ask a few questions about your startup before you use it.
The first question to ask is: “Does your startup have market risk or is it dominated by technical risk?” Lean Startup/Customer Development is used to find answers to the unknowns about customers and markets. Yet some startups such as Biotech don’t have market risk, instead they are dominated by technical risk. This class of startup needs to spend a decade or so proving that the product works, first in a test tube and then in FDA trials. Customer Development is unhelpful here.
The second question is: “What’s the “Market Type” of your startup? Are you entering an existing market, resegmenting an existing market, or creating an entirely new market?” Market Type affects your spending and sales ramp after you reach product/market fit. Startups who burn through their cash, usually fail by not understanding Market Type.
The third question (and the one Eric and I came up with watching the people stream by in Times Square): “What is the “Business Model” of your startup?” Your choice of Business Model affects the metrics you use in discovery and validation and the exit criteria for each step.
Web-based Business Model Exit Criteria
In a web-business model you’re looking for traffic, users, conversion, virality, etc – not revenue. Dave McClure’s AARRR metrics and Andrew Chen‘s specifics on freemium models, viral marketing, user acquisition and engagement both offer examples of exit criteria for Customer Discovery and Validation for startups on the web.
Eric and I will be working on others.