Startup Genome Report Extra On Premature Scaling
Startup Genome Report Extra On Premature Scaling
Startup Genome Report Extra On Premature Scaling
on Premature Scaling
A deep dive into why most high growth startups fail
The goal of this mini-report is to describe our insights about premature scaling,
which we have identified as a major reason high growth technology startups fail.
If you are a startup and you would like to test whether your startup is scaling
prematurely, you sign up for our online web application, the Startup Genome
Compass, at http://startupcompass.co. It is specifically designed to help
startups reduce their chances of failure by benchmarking them to analyze
whether they are scaling prematurely.
Authors
Max Marmer, CSO Startup Genome, Bjoern Lasse Herrmann, CEO Startup Genome,
Ertan Dogrultan, CTO Startup Genome, Ron Berman, PHD at UC Berkeley
Supporters
Chuck Eesley, Stanford University; Steve Blank, Stanford University
Benchmark startupcompass.co
Blog blog.startupcompass.co
Report #1 blog.startupcompass.co/pages/startup-genome-report-1
Report #2 blog.startupcompass.co/pages/startup-genome-report-extra-on-
premature-scal
Methodology www.systemmalfunction.com/2011/08/startup-genome-compass-behind-
scenes.html
Contact [email protected]
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 1
Table of Contents
V. Conclusion .............................................................................................58
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 2
A. The Startup Genome Project and its Role in an
Approaching Societal Transformation
What makes startups succeed or fail? The Startup Genome Project is trying to
answer this question. We believe increasing the success rate of startups has the
potential to dramatically increase economic growth all around the world. On May
28th we released our first report at blog.startupcompass.co. On August 29th we
released our first benchmark application - the Startup Genome Compass to help
startups reduce premature scaling.
The role of technology startups in our global economy has never been more
important. Startups may seem insignificant compared to large multinational
companies that have trillions of dollars of wealth sloshing around in public
markets, but a recent Kauffman Foundation study found that the majority of job
growth in the United States is driven by technology startups.
The power of information technology has been steadily increasing for the last
three decades and has recently reached a level of maturity that has started to
trigger a reorganization of the global economy. It has never been easier or
cheaper to create a startup thanks to infrastructure like open source software,
software as a service, cloud hosting, globally ubiquitous payment processing,
viral distribution channels, real-time collaboration, on demand logistic services
and hyper-targeted advertising.
As a result, the pace of change is speeding up and the implications of this are
immense. Billion dollar startups are emerging faster and faster. The quick ascent
of startups like Google, LinkedIn, Facebook, Twitter, Zynga and Groupon are
harbingers of a major structural economic change on the horizon. The service
sector has dominated the global economy for the last few decades but its sun
will set. Just as machinery replaced most manual labor, software will replace
repetitive intellectual tasks. Turbo Tax eliminated many accountants, Amazon
eliminated many retail jobs and E-Trade eliminated the majority of stockbrokers.
In the near future jobs that are more complex yet still methodical will also be
replaced by software. Creative Commons is reducing the need for lawyers, Khan
Academy shows how one good teacher can replace many bad teachers and the
profession of doctors will be disrupted by startups like Halcyon Molecular that
turn healthcare from emergency care into a preventative self-care. Balancing out
that massive decrease in jobs will be what Richard Florida calls the rise of the
creative class.
As the waves of disruption come ever faster, the only way for a company to be
competitive will be to behave like a startup. In the landmark book the Innovator’s
Dilemma, Clayton Christensen found large companies are excellent at sustaining
innovation but by and large fail at disruptive innovation. Startups thrive on
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 3
creating disruptive innovations. Recently, thought leaders in entrepreneurship
have come to the conclusion that in order for large companies to be effective at
disruptive innovation they need to make structural changes that make them
behave nearly identically to startups.
Eight months ago we launched the Startup Genome Project, with the goal of
increasing the success rate of startups and accelerating the pace of innovation
around the world by turning entrepreneurship into a science. If successful, it's
hard to imagine the type of impact this could have.
Some of the world's biggest transformations occurred when arts were turned
into sciences. The scientific revolution in the 16th century triggered the age of
enlightenment. The development of scientific management, which peaked in the
early 1910’s, made large companies dramatically more efficient and arguably
was one of the biggest causes of the explosion of wealth the world saw in the
last century.
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 4
No revolution is triggered alone. In the quest to make entrepreneurship a
science, we are standing on the shoulders of giants. In just the last 2-3 years the
number of people extracting and codifying the informal learning of entrepreneurs
has hit a point of critical mass. Steve Blank kicked off the move towards a
science of entrepreneurship with his seminal book The Four Steps to the
Epiphany. In the book, he introduced the concept of Customer Development. A
few years later Eric Ries combined Customer Development with Agile
Development and Lean Manufacturing principles to create the Lean Startup
methodology. Interest in the Lean Startup has morphed into a global movement.
Other major contributors to the science of entrepreneurship include Dave
Mcclure on Metrics, Sean Ellis on Marketing, Alex Osterwalder on Business
Models and Paul Graham with his essays.
Yet despite this huge knowledge base emerging about how startups work,
startups have been able to absorb little more than the basic patterns of how to
build a startup. Most founders don't know what they should be focusing on and
consequently dilute their focus or run in the wrong direction. They are regularly
bombarded with advice that seems contradictory, which is often paralyzing. And
while startups are now gathering way more qualitative and quantitative feedback
than they were just a few years ago, their ability to interpret this data and use it
to make better product and business decisions is sorely lacking. The primary
cause of these problems is that we lack the necessary structure to synthesize
our accumulated knowledge on the nature of startups. We are missing a
common language and framework to describe and measure entrepreneurship
and innovation.
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 5
B. Summary of Startup Genome Report
This summary covers the sections of the Startup Genome Report on the different
types and stages of startups. These are important foundational concepts for you
to understand when you reach our section on Premature Scaling in part C. If
you’ve already read the full Startup Genome Report you can skip straight to part
C.
The goal of the Startup Genome report was to lay the foundation for a new
framework for assessing startups more effectively by measuring the thresholds
and milestones of development that Internet startups move through.
Through analyzing the results from our survey we found that Internet startups
move through similar thresholds and milestones of development, which we
segmented into stages. Startups that skipped these stages performed worse.
We also identified three major types of Internet startups with various sub types.
They are segmented based on how they perform customer development and
customer acquisition. Each type has varying behavior regarding factors like time,
skill and money.
These 2 findings lay the foundation for us to begin organizing and structuring all
of a startup’s customer related data, which entrepreneurs can use to make
better product and business decisions.
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 6
Stages. However, in this report only the top level stages are discussed. Our first
four top-level stages are based loosely on Steve Blank's 4 Steps to the
Epiphany, but one key difference is that the Marmer Stages are product centric
rather than company centric.
1) Discovery
2) Validation
3) Efficiency
4) Scale
5) Sustain not covered in this report
6) Conservation not covered in this report
Our assessment of the stages does not include traditional ways of assessment
like funding, team size, user growth, etc. They are based on milestones and
thresholds that vary based on the type of startup. An example for a milestone is
building a minimum viable product. An example for a threshold is certain rate of
retention.
We attempt to provide evidence for the existence of the Marmer Stages in two
ways:
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 7
The Social Transformer / Type 1N
These startups have a self-service customer acquisition strategy and often
create new ways for people to interact. They are almost always confronted with
the challenge of reaching critical mass. If they surpass this threshold they can
often have runaway user growth in a winner-take-all market.
1. Founders that learn are more successful. Startups that have helpful mentors,
track performance metrics effectively, and learn from startup thought leaders
raise 7x more money and have 3.5x better user growth.
2. Startups that pivot once or twice raise 2.5x more money, have 3.6x better
user growth, and are 52% less likely to scale prematurely than startups that
pivot more than 2 times or not at all. A pivot is when a startup decides to change
a major part of its business.
3. Premature scaling is the most common reason for startups to perform worse.
They tend to lose the battle early on by getting ahead of themselves.
Startups can prematurely scale their team, their customer acquisition strategies
or over build the product.
4. Many investors invest 2-3x more capital than necessary in startups in the
discovery phase. They also over-invest in solo founders and founding teams
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 8
without technical cofounders despite indicators that show that these teams have
a much lower probability of success.
5. Hands-on help from investors have little or no effect on the company's
operational performance. But the right mentors significantly influence a
company’s performance and ability to raise money. However, this does not
mean that investors don’t have a significant effect on valuations and M&A.
6. Solo founders take 3.6x longer to reach scale stage compared to a founding
team of 2 and they are 2.3x less likely to pivot.
7. Business-heavy founding teams are 6.2x more likely to successfully scale with
sales-driven startups than with product-centric startups.
8. Technical-heavy founding teams are 3.3x more likely to successfully scale
with product-centric startups without network effects than with product-centric
startups with network effects.
9. Balanced teams with one technical founder and one business founder raise
30% more money, have 2.9x more user growth and are 19% less likely to scale
prematurely than technical or business-heavy founding teams.
10. Founders that don’t work full-time have 4x less user growth and end up
raising 24x less money from investors.
11. Most successful founders are driven by impact rather than experience or
money.
12. 72% of founders find out that their initial intellectual property is not a
competitive advantage.
13. Startups need 2-3 times longer to validate their market than most founders
expect. This underestimation creates the pressure to scale prematurely.
14. Startups that haven’t raised money overestimate their market size by 100x
and often misinterpret their market as new.
15. B2C vs. B2B is not a meaningful segmentation of Internet startups anymore
because the Internet has changed the dynamics of customer interaction. We
found 4 different major groups of startups that all have very different behavior
regarding customer acquisition, time requirements, market risk and team
composition.
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 9
C. Introduction to Premature Scaling
I. Introduction
For the last 6 months the Startup Genome Project has been researching what
makes high growth technology startups successful and has gathered data on
more than 3200 startups. In our research one reason for failure has shown up
again and again: premature scaling.
If you find that definition too abstract, we have a less precise but more concrete
definition, that comes from merging and modifying definitions of a startup from
Steve Blank and Eric Ries:
A startup can maximize its speed of progress by keeping the 5 core dimensions
of a startup Customer, Product, Team, Business Model and Financials in
balance. The art of high growth entrepreneurship is to master the chaos of
getting each of these 5 dimensions to move in time and concert with one
another. Most startup failures can be explained by one or more of these
dimensions falling out of tune with the others.
In our dataset we found that 70% of startups scaled prematurely along some
dimension. While this number seemed high, this may go a long way towards
explaining the 90% failure rate of startups.
The Startup Genome Team trained machine learning algorithms to assess both
the behavioral and actual stage of a startup. If they are not identical the
company has scaled some dimension of the company prematurely.
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 10
II. Definition of inconsistency
• Business Model
• Focusing too much on profit maximization too early
Business • Over-planning, executing without regular feedback loop
Model • Not adapting business model to a changing market
• Failing to focus on the business model and finding out that
you can’t get costs lower than revenue at scale.
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 11
Quote about premature scaling by ...
Michael A. Jackson, serial entrepreneur and investor
“Having been both an entrepreneur and investor I've seen many entrepreneurs
(myself included) try to scale blindly. No one takes venture money to stay a small
business and scaling successfully is what separates eventual industry leaders from
long-forgotten startups in the deadpool. Far too often though entrepreneurs start
scaling before they know what is going to work.
Premature scaling is putting the cart before the proverbial horse, and in the case of
startups this can potentially relate to both engineering and operations. As an
entrepreneur there’s always the temptation to grow the sales team at the first sign
of revenue traction, but there is always the danger that this early traction is coming
from the subset of the market that are early adopters and not the actual market
itself. Additionally, too often I’ve seen startups ramp up sales before they’ve figured
out the most efficient way to achieve profitability. A vicious cycle ensues wherein the
more a company grows, the more it farther away from profitability it becomes.
Teams need to be obsessed with the metrics that drive their businesses’ growth,
constantly testing and challenging their assumptions.
For the engineering team, there’s the often obsessed about notion of having a
robust platform that can handle millions of users before the startup even gets to
10’s of thousands on there. The team starts to worry about all the technical issues
that they’ll have to deal with when success comes, but they lose track of what is
actually going to make them succeed. They can start to fixate on the Engineer’s "El
Dorado" : the Perfect Product. It makes for a nice story that one can dream about,
but it will always remain a fantasy…and the potential cause of opportunity lost.
Venture-backed startups have no option but to scale eventually. Investors have
made their investment based on the fact that they believe the startup is a scalable
business that can attack a large market. Getting venture money can be like putting
a rocket engine on the back of a car. Scaling comes down to making sure the
machine is ready to handle the speed before hitting the accelerator.”
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 12
the market responses and customer behavior. Failing to view those inflection points
may not capture the favorable context for scaling and can be adversely impactful.
Scaling can be analyzed differently based on the context and the circumstances
surrounding a startup and its market. In some instances, a startup can grow to a
certain point and be acquired by a stronger player, thereby avoiding most
premature scaling issues due to lack of resource or experience. However, not every
startup can be quickly acquired, and premature scaling is one of the reasons
startups fail even though they might have a great product, have received funding, or
verified positive market feedback.
Successful scaling for a startup is one of the measures of sustained growth and its
business model, and requires proper preparation and timing. Premature scaling can
happen in situations such as:
• Hiring too many people or paying high salaries before a strong visibility on cash
flow,
• Failing to hire sales oriented people and partners that are skilled, aligned, and
trained to deliver according to the initial patterns of a proven revenue model,
• Confusion between users and customers. In case of consumer focused startups,
when the attention is on scaling users that are not directly impacting profitability.
Many believe that an increase in users or members would somehow translate into
some forms of value.
• Focus on sustained marketing spending without established metrics,
• Early efforts to scale without a strong lock on the market. Market education is
costly for startups beyond early adopters. In addition, other new entrants or more
established companies can benefit from the market being opened by startups
that may not be prepared enough to reap the benefit of their work.
• Startups and their boards need to have mechanisms in place to monitor product
acceptance and market readiness so that they can change gears. Changing
gears does not necessarily mean pivot, but rather being able to capture a larger
portion of the market at a controlled cost/revenue ratio or margin.
• Chasing to increase revenue while decreasing profit. Many startups increase their
costs as they capture additional customers due to lack of fulfillment abilities and
products scaling, therefore putting their business model in jeopardy.”
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 13
III. The anatomy of premature scaling
Startups are temporary organizations that are designed to evolve into large
companies. They move through 6 stages of development throughout their
lifecycle: Discovery, Validation, Efficiency, Scale, Sustain & Conservation. Early
stage startups are designed to search for product/market fit under conditions of
extreme uncertainty. Late stage startups are designed to search for a repeatable
and scalable business model and then scale into large companies designed to
execute under conditions of high certainty.
Every startup has an actual stage and a behavioral stage. Actual stage is
measured by customer response to a product. We measure it by looking at
metrics like numbers of users, user growth, activation rate, retention rate and
revenue. The behavioral stage is made up 5 top level dimensions that the
startup can control. The 5 dimensions are Customer, Product, Team, Financials
and Business Model. Each dimension, both the actual and the 5 behavioral
dimensions are always classified into one of the 6 developmental stages.
A clear example of premature scaling would be a web startup that rapidly scales
up its team to 30-40 people before it has any customers. In this example, the
actual stage of the startup would be in Validation (Stage 2) but the behavioral
stage of the team would be in Scale (Stage 4).
Let's go through some more examples and stats for how each dimension can be
scaled prematurely.
Customer
Stats
Inconsistent startups are 2.3 times more likely to spend more than one standard
deviation above the average on customer acquisition.
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 14
Product
Stats
Inconsistent startups write 3.4 times more lines of code in the discovery phase
and 2.25 times more code in efficiency stage. Inconsistent startup outsource 4-5
times as much of their product development than consistent startups.
In discovery phase 60% of inconsistent startups focus on validating a product
and 80% of consistent startups focus on discovering a problem space. In the
validation phase, where startups should be testing demand for a functional
product, inconsistent startups are 2.2 times more likely to be focused on
streamlining the product and making their customer acquisition process more
efficient than consistent startups. It's widely believed amongst startup thought
leaders, that successful startups succeed because they are good searchers and
failed startups achieve failure by efficiently executing the irrelevant.
Team:
Stats
The team size of startups that scale prematurely is 3 times bigger than the
consistent startups at the same stage. However startups that scale properly end
up having a team size that is 38% bigger at the initial scale stage than
prematurely scaled startups, and almost surely continue to grow. Startups that
scale properly take 76% longer to scale to their team size than startups that
scale prematurely.
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 15
Financials:
Stats
Before scaling, funded inconsistent startups are on average valued twice as
much as consistent startup and raise about three times as much money.
Business Model:
Stats
Inconsistent startups monetize 0.5 to 3 times as many of their customers early
on.
The focus of this post is on premature scaling, but for context, here are a few
example of dysfunctional scaling: Tokbox, Friendster, Orkut, Wesabe, Digg,
SixApart, Myspace (on product), and ChatRoulette.
In our research we also found that the following attributes have no influence on
whether a company is more likely to scale prematurely: market size, product
release cycles, education levels, gender, time that cofounders knew each other,
entrepreneurial experience, age, number of products, type of tools to track
metrics and location.
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 16
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 17
D. Findings on Premature Scaling
I. Summary of the findings
3. Startups that scale properly grow about 20 times faster than startups that
scale prematurely.
4. 93% of startups that scale prematurely never break the $100k revenue per
month threshold.
6. The team size of startups that scale prematurely is 3 times bigger than the
consistent startups at the same stage. However startups that scale properly end
up having a team size that is 38% bigger at the initial scale stage than
prematurely scaled startups, and almost surely continue to grow. Startups that
scale properly take 76% longer to scale to their team size than startups that
scale prematurely.
7. Inconsistent startups are 2.3 times more likely to spend more than one
standard deviation above the average on customer acquisition.
8. Inconsistent startups write 3.4 times more lines of code in the discovery
phase and 2.25 times more code in efficiency stage.
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 18
leaders, that successful startups succeed because they are good searchers and
failed startups achieve failure by efficiently executing the irrelevant.
“Startups are by nature extremely fragile. I'd define premature scaling as expending
money and resources in anticipation of major growth without necessary evidence.
The biggest dangers I've seen in premature scaling are the following areas:
• putting too many features into your product on day 1. At Brickhouse, we'd ask
what are the two features that will have you succeed over everyone else; just build
those. If they work, then add the others. If you have more than 6 major features in
your product, strip it down.
• adding too much management overhead. Once there's a bit of success, it's easy
to go hire the 'rockstar' team. It's critical that the founders keep selling non-stop
and only hire because they can't handle the demand; too often they hire to create
a pipeline rather than responding to existing demand.
• tracking the wrong metric. Often startups look purely at pageviews or unique visits
without looking at the underlying drivers. Tracking metrics that don't fit the business
model is a killer.
• living on the dream. Often, startups are so bought into their own hype they don't
notice that no one is actually buying it. That's great for continued motivation, but it
can lead to real problems. There's a balance between living the dream and the
patience that it'll take time.”
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 19
II. How Consistent Startups Compare to Inconsistent Startup On Key
Customer Related Performance Indicators
20%
0%
0 00 00 00 0+
, 00 ,0 ,0 0,0 0
0- 5
100 0 00 ,00 00,0
50 0 0 - -1, 0-5 5,0
5,0 00 0
0,0 0,0
10 1,0
0
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 20
This could be interpreted ...
Startups that try to scale before they have reached product/market fit and
streamlined their customer acquisition process don't do very well. In fact, no
inconsistent startup in our dataset was able to get more than 100,000 users.
"...Speaking about user base explosion - it is kind of a Holy Grail of the startup
world but if it happens too early and the tsunami of new users attracted by the
network effects will arrive just to find product below their expectations - their will not
come back (eventually at much higher cost). It's better to carefully draft a product
on a smaller, safer group of friendly users then to go after the general public too
soon. The perfect example is Color, which executed PR perfectly, attracted many
users and ... remains a ghost town, as users didn't hook to the product.
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 21
Average user growth per month by stage
10000%
1000%
100%
10%
1%
0%
ry tio
n
nc
y le
ve ca
isc
o
alid
a icie /S
/ D / V / Eff e4
g
g e1 g e2 g e3 Sta
Sta Sta Sta
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 22
Average user growth per 6 month by stage
100000%
10000%
1000%
100%
10%
1%
0%
ry tio
n
nc
y le
ve ca
isc
o
alid
a icie /S
/ D / V / Eff e4
g
g e1 g e2 g e3 Sta
Sta Sta Sta
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 23
Monthly revenue when scaling
50%
25%
0%
0 K k n n n
$ 10 1 00 i llio illio Millio
1K
-
0 k-$ 1M 0 M +
$1 -$ -$1 10
0 0k o n $
$ 1
M illi
$1
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 24
Average valuation for startups that raised money
$ 20000000
$ 15000000
$ 10000000
$ 5000000
$ 0
ery ati
on ncy ale
co
v
id icie / Sc
is al Eff 4
1/
D /V / ge
ge e2 e3 Sta
Sta Stag Stag
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 25
III. Behavior of inconsistent vs. consistent startups
Average teamsize
30.0
22.5
15.0
7.5
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 26
This could be interpreted ...
Startups that scale prematurely have teams that are significantly larger than the
consistent startups that haven't scaled yet, but their team size also rarely gets
as large as startups that do scale properly since they aren't able to sustain their
growth. Scaling the team prematurely is a problem because it's very hard to
align a large team if there are still frequent changes in direction of the company,
as is usually the case in the early stages.
“Great little team with fast product turn around and nimble development that listens
to customer data and does guerilla user testing and makes their product better and
better while seemingly getting into all the best blogs, being loved by opinion leaders
for what they do.
The company's compact team all share a visceral understanding about the product
and hyper growth is not just possible it's expected. The team is two hard core
developers, a designer a UI person who doubles as their tester and a visionary who
has a nose for cheap PR.
They have no cash and no revenues and someone tells them to get a senior
corporate sales guy on board to get some early revenues. He charms the team and
manages to bring on board a few friends who he knows. They all kick in a bit of
cash and start dialing for dollars.
Six months later the company is dead and no one understands why.
The premature scale fail here was not about cost blowouts or technical failures it
happened because the concentrated pool of companies vision was diluted
suddenly by 50 percent with "experience" and "sales feedback"
Scaling the team before the culture and product are really nailed is deadly and
scaling with sales people cuts out the possibility that the product is developed to
sell it's self and both are silent killers.”
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 27
Quote about premature scaling by ...
Krzysztof Kowalczyk, investor and serial entrepreneur
"...In terms of the infrastructure, both technical and HR, premature scaling is
investing too much upfront in building up the capacity which might turn out to be
unnecessary if product/market fit is not be found. It's very important to ensure that
you WILL BE able to scale up quickly once needed, but no more. Look at classic
Bubble1.0 example Webvan, which spent around $1B before understanding their
customers and refining their business model. On the other hand Instagram, a wildly
successful iPhone app, scaled up only when their user base started to grow rapidly.
(http://mashable.com/2011/03/30/scaling-instagram/)."
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 28
Quote about premature scaling by ...
Patrick Paulisch, serial entrepreneur and investor
“Adapting one of my favorite entrepreneurial quotes from race car driver Mario
Andretti: "If everything seems under control, you probably scaled prematurely".
Premature scaling is one of the wrong sides of a fine line any entrepreneur walks on
on a constant basis during the initial phases of a high-growth venture. Anyone not
being on his first venture knows the accompanying signs: on the HR side it starts
with things like spending more than one hour every day in meetings, continues with
the endless discussion of job titles and ends with handing out the pink slips. Things
that help, other than critically reflecting by asking yourself the right questions on a
constant basis: this not being your venture, not buying into your own hype, having
experience on the team and/or board…..or riding on a hockey stick”
“A startup has only a tiny number of ergs of energy. It cannot afford a single minute
or dollar of wasted motion. Pre-mature scaling happens when a startup wastes
valuable ergs of energy on unproven programs and ideas prematurely, before the
winning ideas that are worthy of precious energy are discovered. Startups should
only scale when they know that applying extra energy will create the yields on the
business that they expect. Or as Nathan Furr and Paul Ahlstrom say -- "nail it
before you scale it."”
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 29
Average amount of funding by stage
$ 3000000
$ 2250000
$ 1500000
$ 750000
$ 0
ry tio
n
nc
y le
ve ca
isc
o
alid
a icie /S
/ D / V / Eff e4
g
g e1 g e2 g e3 Sta
Sta Sta Sta
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 30
Startups that raise funding above average - inconsistent vs. consistent
100%
75%
50%
25%
0%
ry tio
n
nc
y le
ve ca
isc
o
alid
a icie /S
/ D / V / Eff e4
g
g e1 g e2 g e3 Sta
Sta Sta Sta
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 31
Quote about premature scaling by ...
Bernard Moon, investor and serial entrepreneur
"If you raise too much capital, especially for inexperienced entrepreneurs, you can
become undisciplined and anxious from the pressure to succeed and grow big.
You might hire too many people, hire too quickly without adequate checks, you
might rush to launch when it might be better to launch late, or you might let inertia
takeover. The last point I mean that you scale too quickly and this could prevent
you from pivoting. Most successful startups change their business model mid-
stream at least once, so premature scaling could blind an entrepreneur from the
company's "real business model".
During my 2nd startup where we raised too much capital, $7 million for our first
round, and proceed to burn through our money like it was doused in gasoline. We
almost went through bankruptcy, by a miracle closed our 2nd round of $7.5 million
during the bust of 2000, changed the business model (premature scaling at it's
best), and then became a profitable company."
“Spending money beyond the essentials on growing the business (e.g., hiring
sales personnel, expensive marketing, perfecting the product, leasing offices,
etc.) before nailing the product/market fit.”
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 32
Money spent on customer acquisition per month before funnels & costs optimization
40%
20%
0%
0 00 00 00 0+
$ 10 2,0 ,0 ,0 ,00
$0
- -$ -$15 00 00
0 0
00 -$1 $1
$1 2,0 00
$ 5,0
$1
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 33
When startups spend this much money on customer acquisition, they are trying
to scale, the problem is it's not very effective and leads to inconsistency if
startups haven't first built a product customers want and can efficiently acquire
customers.
Premature scaling is indeed a frequent cause of failure for web-related start ups. I
have witnessed it several times.
1) Definition
For me premature scaling refers to a situation where money is spent towards
growing the business before a viable business model has been identified. Until that
point has been reached, money should only be used to test a wide array of options
to determine which model might work. Variations in product features, claimed
benefits, distribution channels, customer targets, partnership arrangements, etc.,
result in a huge number of combinations, most of which do not constitute viable
businesses. Unless many of these are tested, the initial model is likely to be way
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 34
under-optimized. Betting on growth first, and 'refining' the model along the way is
exceedingly dangerous because the alternatives are often incompatible strategies,
not just minor refinements. Having committed a lot of cash - and therefore
reputation - to a course of action makes it very difficult to change course. By the
time various operational adjustments have failed to resolve the problem, the
company may run out of money and, most damning, credibility.
2) Example
This french company introduced the first 100% on-line gift-cheque (as opposed to
paper coupons, with a complex logistics attached). The business concept looks
wonderful and eminently scalable: it consists of literally issuing money at zero
variable cost; all the operating cost consists of running the web platform. The big
issue is to reach scale to satisfy both vendors and users, and of course cover the
fixed costs.
A 300k € seed round allowed to build the platform, garner a few vendors and start
operations. The next step was to sign up a range of vendors to provide a wide
enough choice of products, and to draw customers in large enough numbers to
satisfy the vendors' expectations. A substantial additional round was completed to
launch a full scale marketing campaign, mostly aimed at attracting individual
customers to the site. Because a disproportionate share of the business takes
place in the Xmas season, it took almost a year to ascertain that the direct
approach to individual customers didn't work. For the following season, the focus
shifted to partnering with existing communities of potential users, some of which
already using paper gift-cheques. Another round was raised to bankroll the new
strategy. The results were much better, but still not good enough to reach self-
financing. By that time, the main investor had lost faith in the business and the
team. The company was sold to one of the direct competitors.
This business doesn't easily lend itself to quick, low cost testing, but, with
hindsight, I can see how we should have reversed the approach: first go for very
narrowly defined deals (in partnership, with focused targets), and based on the
learning, move gradually to larger segments and a more autonomous model. This
could have been done with a much reduced team, and a ratio of revenue/marketing
cost that would have justified sustained investing until break-even.
“Premature scaling is when a baby tries to run before it knows how to walk. In other
words, spending excessive amounts of resources to expand upon activities that are
fundamentally flawed. However, at some point entrepreneurs need to attempt to
scale to learn. Therefore, premature scaling may mean lack of intention around
systematic scaling experiments to validate assumptions.”
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 35
Self Reported Valuation of startups that have not raised funding
$ 20000000
$ 10000000
$ 0
Stage 1 / Discovery Stage 2 / Validation
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 36
Lines of code written by stage
400000
300000
200000
100000
0
ry tio
n
nc
y ale
ve c
isc
o
alid
a icie /S
/ D / V / Eff e4
g
e1 e2 g e3 Sta
Sta
g
Stag Sta
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 37
Quote about premature scaling by ...
Alex Barrera, Serial Entrepreneur
"Premature scaling is trying to fix a problem you haven't encountered yet. It tends to
happen mainly to tech founders where their desire to build cool technology gets in
front of the real problem solving. You can detect it in long release cycles, constant
failure to deliver and unused capabilities and/or product features. Premature scaling
is probably the main reason why most startups end up spending all their cash (and
dying). One of my own startups incurred that and was fatal, but I know much bigger
startups that raised millions just to spend them hiring hardcore backend engineers,
building a robust and super scalable backend just cause their founders user goals
were X. What happen? Well, they spent all their cash, wasted 1 year and got 0
users."
“1) Premature optimization - selecting tools or technology platforms that are slower
to develop on / not as-familiar to the team but theoretically higher-performance in
anticipation of potentially-large demand but in reality prevent there from being large
demand since the feature set of the product is so slow/hard to work on.
We have seen our fair share of premature scaling over the years.
Some in our own portfolio, some in companies we were looking at and (luckily)
decided to pass on.
If I look at these last 10 years, the one regular occurrence of the “cart before the
horses” syndrome was in the wrong estimate or market demand timing. As some
folks have already written, demand is not so easy to estimate when you are there.
After the fact, it does look simple, but it is not. The problem with premature scaling
was exacerbated by fixed nature of the technology costs (culprit #1). Before AWS
you could only buy technology with a step function cost structure. Some tech-led
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 38
companies with too much money (culprit #2) decided to install the latest Unix
servers to test beta customers. Other companies were mistaking a handful of
early adopter clients in a new country as the compulsory call to build a new full
infrastructure there. Then, at the first macroeconomic shake up (and we have
one every 2 years it seems) the new territory still has less than 10 clients, and
more than 10 staff. Unwinding a prematurely scaled organization may be more
expensive than building it in the first place. Back when Jeffrey Moore gospel
was new, a few companies (and VCs) misread the whole Gorilla theory. Get big
first, get the biggest market share and as the market leader reap the big
benefits. The serious fallacy in that argument was its assumption that market
demand is there now. And it assumes that demand will be there tomorrow.
I actually think that cloud computing and open source software and an ever
decreasing cost of bandwidth are contributing to help start ups from premature
scaling, at least on the area of fixed technology costs. The ability to tune down
the power, to test and pivot twice or more on the same seed funding dollar can
multiply the chances of success.
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 39
Focus of the team in discovery
50%
25%
0%
nt /50 en
t
pme 50 m
l o lop
e ve e ve
rD tD
me d uc
to Pro
C us
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 40
Self assesed priority in discovery stage
80%
40%
0%
t t t
blem d uc d uc aling a t ion duc
pro g pro g pro Sc im
iz
g pro
n g i n i n a x i n
eri lid
at
ml
in it m sif
y
c ov a a o f e r
Dis
V
Str
e Pr Div
35%
0%
t t n t
blem d uc d uc aling a tio duc
ro ro ro S c i z ro
in gp i n gp i n gp a xim i ngp
er a t in m fy
c ov a lid a ml ro fit e rsi
V e Di v
Dis Str P
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 41
Product stage before product/market fit
40%
30%
20%
10%
0%
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 42
Quote about premature scaling by ...
Mick Liubinskas, serial entrepreneur and co-founder of Pollenizer
“Scaling at the right time is tough. Too early and you waste money and get
distracted. Too late and you miss the market or run out of runway.
It's not scientific but the number we talk about at Pollenizer is 60%. If you've got
your business model working to 60% of your projections then go for it. It means
you're past half way there but there is still plenty of room for optimization. We're
after confidence without perfection.
Don't think about scaling as purely about more customers. One of the most
dangerous ways you can scale is through new features. Every feature you add is
something else you need to maintain, manage, measure, market, understand and
support. The same applies to markets. Too many segments and you're juggling too
many needs with too few resources and hours in the day.
The main thing is for you and your team to choose the right time. Don't be under
pressure from investors, media or even customers. Right or wrong, you need to
trust your instincts.”
“Founders often forget that approximately >200k early adopters exist around the
globe. Their goal is to try every new product - but not to use or pay for it. Therefore
200K free users mean very little. Instead of scaling your user base prematurely
good startups are focusing on small defined user groups. Their goal is to
understand how their product/service offering fit well with this defined user group,
to estimate the size of this group, to discover the key triggers that turn the
members of this group into real customer, to learn what features need to be added
to the product/service keeping on mind to win the hearts of customer segments
beyond their original user group.
(…)
In modern time, you can't cultivate your start up in a very predicted "agriculture"
order. Market condition and requirements are changing so often that you need to
live the " jungle" culture. You need to react and balance your strategy more
dynamically. Since not the best and strong trees are surviving in the jungle, but
more flexible and adoptive.”
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 43
Average percentage of team outsorced before product/market fit
20%
10%
0%
Stage 1 / Discovery Stage 2 / Validation
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 44
Type of market
50%
25%
0%
tter) pe
r)
ich
e) rk et
(be he
a
t (n Ma
et t (c rk e w
rk e
ma Ne
g ma ark g
tin m tin
is t i ng is
E x is Ex
Ex
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 45
Percentage of paid users by type before validation
40%
20%
0%
e1 1N e2 e3
Typ yp
e yp yp
/ /T /T /T
a tor e r
rat
or
n ge
r
tom form t e g alle
Au r a ns In Ch
lT
cia
So
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 46
Distribution of startups by Type
70%
35%
0%
e1 1N e2 e3
Typ Typ
e Typ Typ
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 47
IV. Similarities of inconsistent vs. consistent startups
The last section contained lots of data on how consistent companies behave
differently from inconsistent companies. This sections contains data to show the
many ways that they are very similar to each other, which highlights the
importance of the conclusions in the previous sections.
20%
0%
n n n n n n n
llio Bil
lio
Bil
lio l lio l lio l lio l lio
Mi B i B i B i B i
1 00 - $1 $10 $50 100 5 00 5 00
- - $ $ $
<$ illio
n
lio
n
lio
n
50
-
00
- >
M B i l B il $ $ 1
00 $1 $1
0
$1
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 48
Frequency for product release cycle
30%
15%
0%
y y k k ths s ths ths
r da r da w ee w ee o n e ek o n o n
e e r r w
e s p ce p s pe e pe
1 - 3m 2 -3
3 - 6m n 6m
tim O n e c ry
le tim On v ery E ve v ery e tha
i p -3 E E r
M ult 2 Mo
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 49
Level of education of founders
50%
25%
0%
ne l e l e e
oo eg oo gre gre
No ch oll Sc
h e e
hS e C
ate te
D
te
D
Hig So
m u u a u a
G rad rg rad G rad
me de
So Un
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 50
Gender of founders
100%
50%
0%
Female Male
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 51
Time co-founders have known each other before starting
50%
25%
0%
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 52
Number of technology startups the founders have founded before
80%
60%
40%
20%
0%
0 1 2 3 4 5 6+
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 53
Average age of the founders
60 Years
39 Years
35 Years 34 Years
18 Years
Average Age
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 54
Number of products
80%
40%
0%
1 2 3+
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 55
Top 3 tools to track metrics
80%
40%
0%
Google Analytics Homegrown Solution Spreadsheet
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 56
Distribution by country
60%
30%
0%
ca pe ia ia t ica ca
ri As an as ri
me Eu
ro e E Afr me
hA Oc d dle A
rt Mi uth
No So
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 57
V. Conclusion
Concluding we can say that we were surprised with the clarity of the results.
Since our last report we have grown our dataset to more than 3200 firms and
completely overhauled our classification methodology with machine learning
algorithms.
Our algorithms classify each firm into a behavioral stage and an actual stage
according to the Marmer stages model. Using this classification, we are able to
recognize firms whose behavioral stage does not match the actual stage. These
firms are called inconsistent, and their main problem is that they scale their
operations prematurely.
The issue with firms that scale prematurely is that they perform significantly
worse than firms that scale properly. Inconsistent firms have lower user growth,
raise less money in later stages and make less revenue per month.
A notable finding in our research is that so many firms were prone to premature
scaling. Specifically, 70% of the firms in our dataset exhibited such behavior. We
are able to show that many commonly quoted causes for failure are probably not
the causes of premature scaling. The age of entrepreneurs, their location, their
experience nor their education affect their chances of scaling prematurely.
The work is far from complete. Two future research topics currently seem
promising:
Improving our machine learning algorithms to classify and analyze data more
accurately, and make use of new firm data being collected.
Longitudinal data of firms and its analysis will allow firms to extrapolate
individual recommendations on how to operate in different environments.
In order to aid firms in avoiding the pitfalls of early scaling, we developed the
Startup Genome Compass. The Compass uses firm data to generate a
comprehensive report for each individual firm. The report contains benchmarks
and analyses that can be used by entrepreneurs to ascertain their startup Type
and Stage. We hope these insights will help firms identify issues that require
their attention when advancing through the Startup Lifecycle.
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 58
D. Acknowledgments & Sources
Authors:
Max Marmer, Bjoern Lasse Herrmann, Ron Berman, Ertan Dogrultan
Supporters:
Chuck Eesley, Steve Blank, Aleksandra Markova
Methodology:
In order to not bias the startups taking the survey we will not publish the
thresholds and milestones we use for the stage assessment. More info can be
found here http://www.systemmalfunction.com.
Underlying Assumptions:
• We assume that a company is tackling a large market from the get go or
navigates towards a large market as they evolve
• We assume that the company is able to place themselves in an ecosystem
where talent and money is accessible
• This view makes sense predominantly for companies with high market risk.
We assume a startup is able to surmount their technology risk. Companies
with large technology risk may fail for additional reasons.
Sample sizes:
• Total number of startups: 3200
• Startups that were consistent with the Marmer stages: 30%
• Startups that raised $50k+ funding (with specific amount disclosed): 316
• Startups that had funding but not disclose the amount: 392
• All startups were high growth technology startups
Inspiration:
Steve Blank, Dave McClure, Sean Ellis, Eric Ries, Paul Graham, Joel York
Startups that scale prematurely are classified as inconsistent and startups that scale properly are classified as consistent
Startup Genome Report: premature scaling v 1.2 (edited March 2012) . Copyright 2011 . Page 59