04 Atlassian Case Study Discussion PDF

Atlassian Case Study Answers and Discussion
1. If you were T. Rowe Price, would you invest $150 million USD in Atlassian at a
valuation of $3.3 billion USD? Why or why not?
If we were T. Rowe Price, we would NOT do this deal because it is exceptionally unlikely that
well be able to realize our targeted 20% IRR.
While Atlassian is a great company, the $3.3 billion valuation, along with its expected growth
rates and margins, make a 20% IRR over a 3-5 year holding period implausible.
The companys valuation at the time of the T. Rowe Price investment approximately 110x
EBITDA was already exceptionally high.
To achieve a 20% IRR, however, the companys value would need to climb even higher to the
130-140x range in our Base Case scenario if there is no liquidity event for another 3-4 years:
While this is possible, generally the multiples decline over time as the company grows.
Here, Atlassian grows from $200 million in revenue (34% historical growth) in FY14 to $459
million (12% trailing growth) in FY19.
It is extremely unlikely that a $459 million company growing at 12% per year will be valued at a
higher multiple than a $200 million company growing at 34% per year.
In our Base Case scenario, the customer count grows from 35,000 to 70,000 over 5-years (15%
CAGR), while the average selling price increases by 3% per year for 5 years. We assume that the
operating margin falls from 10.6% to 8.6% over 5 years, primarily because of added sales &
marketing spending.
Even with more optimistic assumptions, however, it is highly unlikely that T. Rowe Price will
realize a 20% IRR.
For example, if the customer count increased to 100,000 rather than 70,000, implying Year 5
revenue of $645 million, the baseline IRR would only increase marginally, from 13.8% to 15.0%:
http://breakingintowallstreet.com
http://www.mergersandinquisitions.com
This happens because the average sales & marketing expense per new customer increases from
approximately $2,000 in FY14 to $4,000 in FY19 (in the 5 years prior to that, it increased from
$1,500 to $2,000).
This large increase reflects our view that it will become more and more expensive to win
customer accounts at bigger companies with higher user counts.
If the average sales & marketing expense per new customer only increased to $3,000 over 5
years, we would achieve exactly a 20% IRR over the 5-year holding period, still assuming
100,000 customers and the same exit multiple:
But there is still a problem even under these optimistic assumptions: the exit multiple must
stay the same for T. Rowe Price to realize a 20% IRR:
If the multiple drops by even ~10%, the IRRs fall to below 20%; and if it drops by more than
that, the IRRs fall even more substantially.
So to achieve a 20% IRR:
The companys customer base has to triple;

Its average sales & marketing expense per new customer also has to increase by 50%
less than expected; and
The exit multiple has to remain the same even though Atlassians growth rate and
margins are both lower by Year 5.
It is extremely unlikely that all of these will happen, but it is possible that one or two of them
may transpire over the next 5 years.
While Atlassian is a great company, we see it as a 10-15% IRR investment as opposed to one in
the 20%+ range, based on the different scenarios we have analyzed.
As a result, we recommend AGAINST investing $150 million in the company at a valuation of
$3.3 billion.
2. If your answer is no, what conditions would have to be true for you to change your
decision and invest in the company anyway? For example, if the companys growth
rates, margins, or its valuation were different, would your decision change?
If your answer is yes, what investment terms might you stipulate to protect yourself
in this deal?
The greatest uncertainty here surrounds Atlassians margins and sales & marketing spending in
the future.
As shown above, the customer count and revenue growth rate make far less of a difference
than the margins do.
Even if the customer count increased to 140,000 over 5 years a 4x increase the IRR would
still not reach 20% under our baseline assumptions!
We also view an increase from 35,000 to 140,000 customers over 5 years as completely
implausible, given that the customer base only increased from 17,000 to 35,000 over the past 4
years.
Therefore, a condition on the customer count is not particularly realistic or helpful.
Instead, we would look to the companys sales & marketing spending and its pricing policies as
better levers to pull.
On the margin / expense side, suppose that we used the following assumptions instead (while
keeping the total customer count at 70,000 by Year 5):
Sales & Marketing per New Customer: Increases by only 4% per year over 5 years,
reaching $2,500 by the end instead of $4,000.
Sales & Marketing per Existing Customer: Increases by only 4% per year over 5 years,
reaching $1,150 by the end instead of $1,400.
G&A Variable Expenses: It falls from 10% of revenue to 8% over 5 years, rather than
increasing from 10% of revenue to 12%.
With this set of assumptions, the companys revenue still increases to $459 million after 5
years, but now the EBITDA margin increases from 16% to 22% in that same period:
And the IRR also increases so much that even with a ~20% drop in valuation, a 5-year 20% IRR is
still plausible:
Of course, there are other ways to get to the same outcome.

For example, the company could gradually scale down its R&D spending from 40% of revenue
to 30% of revenue over the 5-year period, and achieve a final year EBITDA margin of 23%:
It is unclear if this substantial cut in R&D spending will impact the companys sales growth at all.
Thirty percent of revenue spent on R&D is still quite significant, but we would need more detail
on the deployment of technical and engineering employees before we could determine if this
reduction is feasible.
Finally, it is also worth asking if Atlassian has room to increase its product pricing, rather than
reducing its pricing as it has done in prior years.
Our baseline scenario assumes only a 3% increase in average customer value per year,
reflecting modest price increases to keep pace with inflation.
But if the average customer value increased by 5% per year instead, the 5-year IRR would
increase to 20% as well.
Although the companys revenue does not increase by a huge amount (in the final year its $505
million rather than $459 million), its EBITDA margin stays at ~16% instead of declining to ~14%,
which makes a big difference in the final numbers:
And there may be room to increase prices by even more than 5% per year even if customer
growth slows as a result of these price increases, the trade-off could easily be worth it,
especially if new customer acquisition costs increase as we have predicted.
In summary, the following conditions would have to be true for us to feel comfortable investing
$150 million in Atlassian at a valuation of $3.3 billion:
The EBITDA Margin would have to increase from 16% to at least 20% over the 5-year
projection period, via some combination of lower-than-expected increases in customer
acquisition costs, reduced R&D spending, and/or reduced G&A spending; and
Average customer value would have to increase by at least 4-5% per year rather than
3% per year over the 5-year projection period in this case study.
3. At an Atlassian valuation of $3.3 billion, Accels 15% stake is worth $495 million, so it
was not possible for them to sell the entire stake to T. Rowe Price. However, press
releases and news reports confirm that Accel did sell at least some of its stake, taking
money off the table.
Do you agree with this decision? How much do you think Accel should have sold, and
why? Or if you disagree with the decision, why would Accel be better off waiting to
sell?
Since Accel is targeting a 10% money-on-money (MoM) multiple with its investment, we need
to look at the multiple under different scenarios to answer this question.
In our baseline scenario, with revenue growing to $459 million and final year EBITDA of $61.8
million (14% margin), Accel seems likely to realize a multiple well above 10x if Atlassians
valuation stays the same or increases over the next 5 years:
Even if Atlassians valuation falls by over 40%, Accel will still realize a 10x multiple on its
investment if it waits until FY19 to sell its stake.
So on the surface, the firms decision to sell part of its stake does not make much sense:
AT MOST, Accel could sell $150 million of its stake to T. Rowe Price, leaving it with a
stake worth $345 million, or 10.5% of the company.
$150 million represents a 2.5x return on Accels initial investment of $60 million, which
is a long ways away from its stated goal of a 10x return.
In some sense, this reduces the investment risk because now Accel has taken money off
the table and recovered 2.5x of its initial investment. So even if Atlassian collapses or
otherwise turns into a disaster, Accel cannot possibly lose money on this deal.
But with a company like Atlassian it is exceptionally unlikely that returns would ever turn
negative because it is still growing quickly and actually commanded a higher multiple 4
years after Accels initial investment.
And owning LESS of a company only limits investment losses in extreme downside
scenarios where the returns turn negative.
To illustrate these points, lets take a look at what happens if we assume that the entire $150
million investment from T. Rowe Price goes to Accel cashing out part of its existing stake:
While the IRR increases in this scenario, the money-on-money multiple actually decreases since
Accel owns 4.5% less of the company by the end of the 5-year projection period:
Of course, Accel did not actually sell $150 million of its stake to T. Rowe Price since most of the
shares came from employees selling their stock.
So if Accel sold only $20 million or $50 million, its potential money-on-money multiples would
not decrease by as much as what was shown here but they would still decline.
This move to sell shares would make more sense for a later-stage investor that really IS looking
to take money off the table and is aiming for only a 2-3x return on investment.
In short, Accels move to sell some of its stake does not make much sense because it actually
hurts its chances to realize a 10x return on its initial investment.
This move would only make sense if:
Accel has detailed knowledge that the companys future financial performance will be
significantly worse than what we have projected here (thereby reducing final year
EBITDA and the likely EBITDA exit multiple).
It is actually targeting a specific IRR rather than a money-on-money multiple.
Accel knows or believes that Atlassian will not sell or go public for a very long time
(more than 5 years in the future), and it wants to realize some of its returns earlier than
that.
4. If you were to pursue this investment opportunity, what areas would you need to
perform further due diligence on? For example, are there any assumptions critical to
your view of the company that you had less conviction on? What are the top 2-3 most
important figures and assumptions that you need to research in more detail?
As mentioned above, the key drivers in this deal are the companys planned sales & marketing
spending, R&D spending, and pricing increases.
To a lesser extent, its customer growth expectations and total potential customer base also
make a difference, and we would need to perform further due diligence on those to feel 100%
comfortable with this deal.
The top 2-3 most important figures and assumptions we need to research in more detail are as
follows:
Customer Acquisition Costs: What is the historical trend for Sales & Marketing Spending
per New Customer and Sales & Marketing Spending per Existing Customer, and how
does the company expect both these numbers to change in the future? This is critical
because new customers may actually reduce the companys margins if customer
acquisition costs grow too high.
Plans for R&D Spending: Atlassians business model allows it to spend significantly more
on research & development than peer companies, but this is both a strength and a
weakness since 40% of sales spent on R&D is quite high. What are the companys future
plans, and how important are these activities to revenue growth? Even a modest
reduction in R&D spending from 40% to 35% of revenue could make a huge impact on
the returns.
Average Customer Value / Plans for Product Pricing: Historically, the company has
reduced pricing on many of its products. This may not have been the best decision,
because its average customer value has not changed much over the past 4 years while
its per-customer sales & marketing expenses have increased
If there is room to increase average customer value by more than 3% per year, that
could also significantly increase the IRR; but if there is substantial downward pricing
pressure, that might force the companys margins down and further reduce IRR.
Other interesting, but less important, areas include:
Total Addressable Market and Relationship to Sales & Marketing: We assume in the
baseline scenario that Atlassian increases its customer count to 70,000 at the end of 5
years but is that plausible?
Atlassians pricing for the OnDemand version of JIRA, as of the time of this case study,
was $500 / month for 500 users and $300 / month for 100 users. The Download version
pricing was $4,000 for 100 users and $8,000 for 500 users.
Since Atlassians average customer value is close to $6,000, this implies that its average
customer has close to 500 users.
But Atlassian sells tools primarily for software developers and engineering organizations
which brings up the question of how many companies worldwide have hundreds of
employees in that function. There may not actually be 100,000, or even 70,000,
potential customers in this market.
It would also be interesting to know how much additional sales & marketing spending is
required to reach these new customers, and whether or not the company will need
commissioned sales reps to grow past a certain point.
OnDemand vs. Download Customer Split: In our Base Case scenario, we have not
assumed a massive shift to the OnDemand segment about 15% of customers are
OnDemand in FY14 vs. 19% in FY19.
However, this could easily be wrong and that would have implications for the average
customer value, customer acquisition costs, and more.
In particular, it would be interesting to know the sales & marketing spending per
OnDemand customer vs. Download customer, and see how that compares to the sales
& marketing spending per existing customer vs. new customer.
G&A Spending: While this is less significant than sales & marketing, it does still comprise
around 13% of revenue in Atlassians most recent fiscal year, and its an area ripe for
potential cost savings.
Cutting sales & marketing spending may not even be possible, and cutting R&D spending
might have negative long-term consequences, but general & administrative costs can
often be reduced with less hassle.
5. In hindsight, was it a good idea for Atlassian to reduce its prices significantly in 2011 to
win more customers, and to offer the on-demand model for its products? What are
the financial implications of these decisions?
It is tough to answer this question because we dont know the counter-factual how
Atlassians customer count would have changed WITHOUT the reduced pricing.
However, we can make an educated guess about the financial and operational impact of these
decisions:
FY09: 10,000 customers; revenue of $44 million; average customer value of $4,426.
For reference, Atlassian changed its JIRA pricing from $150 / month for 10 users to $10 / month
for 10 users in FY11.
Technically, it seems like this reduced pricing greatly reduced the average customer value.
But the FY11 value was already a substantial increase over the average customer value in the
two prior years, so it may not mean that much.
On the other hand, it also doesnt seem like this pricing change greatly increased Atlassians
customer count.
By reducing the per-user prices, Atlassian shifted its business model and must now focus on
acquiring more users from existing customers to make up the difference.
Given that it is significantly more expensive to win new customers than to market to existing
ones, this may have been a wise decision if Atlassian is confident it can boost the user count at
existing accounts.
However, it also means that the company must focus more on large accounts with a high
number of potential users if customer acquisition costs are currently ~$2,000, it no longer
makes sense to pursue companies with only 10 potential users.
Bottom-Line: We would probably rate this a neutral impact in the absence of further
information.
The switch to on-demand is even harder to assess because we need more information on the
average number of years customers stay subscribed, the churn rate (the percentage that cancel
each year), and the breakeven point at which new customers become profitable.
Broadly speaking, switching to subscription-based business models hurts accounting profits
(Net Income) in the short-term because revenue must be recognized over a long time period,
but many costs such as commissions are paid upfront but in the long-term, it can actually
boost cash flow due to the high amount of deferred revenue being generated.
We expect a more positive impact on Atlassian because the company does not have
commissioned sales reps, which eliminates one of the biggest upfront costs so the profit
impact is likely neutral, with a positive long-term cash flow impact.
6. Do you think the growth numbers and margins you have projected are achievable?
What is the relationship between average customer value, growth, and the total
number of customers the company can potentially sell to?
Yes, the growth and margin numbers here seem achievable because revenue growth rates
decline each year and are well below historical levels, the total customer count by the end of
the period does not seem outrageous, and the margins actually decrease by 2-3% due to
additional S&M and G&A spending.
Revenue: Historically, revenue doubled between FY11 and FY14, growing from $102
million to $200 million; here, it also doubles over 4 years, growing from $200 million in
FY14 to $412 million in FY18.
Revenue Growth Rates: These decline over time, starting at 25% in FY15 and then
falling to 12% by FY19.
Source of Growth: Most revenue growth comes from winning additional customers, as
opposed to pricing increases, with the count growing from 35,000 to 70,000+ by FY19. In
the 4 years prior to this, the customer base increased from 17,000 to 35,000.
Margins: Historically, the companys Operating Margin was in the 7-11% range; here, it
starts at 11% and falls to 9% by the end. The EBITDA margin was 12-17% historically, and
here it falls from 16% to 14% by the end.
Expenses: Some expenses, such as G&A, actually increase as percentages of revenue; for
others, such as sales & marketing, we increase the per-customer expense substantially
over time in order to be as conservative as possible.
There is a question over the total addressable market, and whether or not the company can
really serve 70,000+ customers eventually.
We believe that to reach that level, Atlassian will have to spend significantly more on sales &
marketing since many of these customers are larger accounts that require more time and effort
to sell to.
7. How do Atlassians business and business model compare to other hot tech
companies at this point in time, such as Box, Dropbox, and Square? What is more
appealing and what is less appealing vis--vis peer companies? What about compared
to companies in similar markets, such as 37 Signals / Basecamp, Wrike, and Github?
Atlassian is far more profitable and cash flow-positive than companies such as Box, Dropbox,
and Square, most of which are losing enormous sums of money.
That alone makes it a more appealing business, but there are several other factors to consider.
For one, Atlassian sells to very different users and customers than these other companies it is
focused on developers and more technical employees, while the others offer more general
purpose tools and software for businesses at large.
The companies mentioned in the second set above Basecamp, Wrike, and Github are much
smaller and/or newer, so we focus on Box, Dropbox, and Square in the comparison below:
Growth and Growth Potential: Box, Dropbox, and Square have grown at faster rates
than Atlassian in recent years, and may have higher future growth potential.
According to its S-1 filing, for example, Box grew from $59 million in revenue to $124
million in revenue in its most recent fiscal year as of the time of this case study,
representing 110% growth, compared to Atlassians 34% growth.
Dropbox, meanwhile, grew sales from $46 million in 2011 to $116 million in 2012 (152%
growth) to approximately $200 million in 2013 (72% growth).
As of the time of this case study, Square had disclosed less information about its
financials, but its net revenue was approximately $130-140 million in 2013, with
revenue expected to double in 2014 ($260-280 million), also representing significantly
higher growth than Atlassian.
Furthermore, these 3 companies all serve broader markets (online file storage/sharing
and payment processing) than Atlassian, so it is likely they will continue to grow more
quickly in the future.
Profits and Cash Flow: Here, Atlassian comes out ahead: it has been profitable and cash
flow-positive since inception, with a Net Income Margin of 7% in FY14, an EBITDA
margin of 16%, and an FCF margin of 15%.
By contrast, most other peer companies were losing massive amounts of money: Box,
for example, lost $158 million in its most recent fiscal year and spent 138% of its
revenue on sales & marketing (vs. ~21% for Atlassian).
Squares numbers are undisclosed, but sources indicate that the company lost $100
million in 2013 (on revenue of $130-140 million).
Dropbox has allegedly been profitable, but it is unclear what its margins and whether or
not it is cash flow-positive in addition to being profitable.
If it really is profitable, it is likely because it is spending less on sales & marketing as a

percentage of revenue than other software companies and still has relatively high gross
margins.
Threat of New Entrants: These markets are all highly competitive, with new entrants
both large and small emerging quite frequently.
In online file storage, for example, huge companies such as Apple, Microsoft, Google,
and Amazon have been introducing their own products and services since they do not
require huge upfront R&D investments (compared to something like Googles search
algorithm, online file storage is a relatively simple technical problem).
The payment processing space (Square), similarly, has many competitors both large
(Paypal) and small (other startups) because there isnt much loyalty to particular
vendors and its also not an extremely difficult technical problem.
In some ways, Atlassian is similar to these other markets: the technology itself is not
that complex because source code control and collaboration tools have been around for
decades.
The difference is that there are not quite as many large companies attempting to enter
the market, likely because its much smaller and more fragmented.
Threat of Substitute Products/Services: There is a high threat of substitute

products/services for almost all these companies, because theres little to force
customers to stick with their existing solutions.
For example, a large company that signed up for file storage via Dropbox or Box could
easily switch to a competitive service offered by Microsoft or Google all they have to
do is copy and paste their files into a new service and wait for the upload to complete.
The same applies to payment processing (Square), but there it is even worse because
small businesses and micro-merchants do not care much about payment processing
and tend to use whichever solution is easiest and cheapest.
As a result, many of these companies have been attempting to move up-market and
offer other business software that is stickier.
The same is true of Atlassian: there are so many source code control and project
management / collaboration tools that very little prevents customers from switching to
solutions offered by companies like 37 Signals / Basecamp, Wrike, or Github.
Github probably has the lowest threat of substitute products/services because of the
social component: it is actually more of a social network for developers than a true
enterprise software company.
Since users can view other users contributions and review source code from many
different companies and individuals, over time more and more users are likely to use
Github and stick with it since everyone else is using it as well.
Bargaining Power of Customers (Buyers): In most of these markets, there is substantial

downward pricing pressure in online file storage, for example, Google shook up the
industry around the time of this case study by dramatically reducing its Google Drive
prices (10 TB used to be $800 / month, but they reduced prices to $100 / month).
Since there are so many substitutes for the products/services offered by these
companies, and since the solutions themselves are not highly differentiated, customers
have a great deal of bargaining power.
This is also true of Atlassian, as seen with its move to dramatically reduce prices several
years ago; it isnt exactly in a price war with Github because the two companies price
their services differently (# of repositories vs. # of users), but it is unclear how much of a
premium Atlassian can actually charge.
The company could gain additional bargaining power if its software were more
technically complex, if there were more of a lock-in effect, or if it offered even more
specialized solutions (e.g., different versions of its tools for different industries).
Bargaining Power of Suppliers: This point is less relevant for software and Internet
companies because there are no suppliers in the traditional sense, but it is worth
looking at labor costs since the employees are arguably the suppliers here.
Companies like Box, Dropbox, and Square compete with each other for the top
engineers and sales people, and often pay high base salaries even to entry-level
employees (e.g., $100K+ for engineers just out of undergraduate). They have little
bargaining power because engineers could easily leave and move to other startups or
large tech companies offering the same, or higher, salaries.
Atlassian faces some of the same issues, but since it is based in Australia it arguably has
more power over employee wages: there are not as many tech companies or promising
employers for talented engineers and sales people there, so there may be less upward
pressure on wages. As the company expands overseas, however, this advantage may
diminish over time.
Intensity of Competitive Rivalry: As mentioned above, these markets (online file

storage, payment processing, and developer/collaboration tools) are all highly
competitive, with relatively low barriers to entry, frequent new entrants, and little
product differentiation.
Atlassians market, in particular, has dozens of competitors offering slightly different

collaboration/project management/source code control tools, ranging from high-profile
venture-backed companies (Asana, Github) to self-funded companies that have been in
business for 15+ years (Basecamp / 37 Signals).
Online commentary reveals that there is no clear product winner, and that the best
solution comes down to user preferences and individual requirements.
8. Even under conservative assumptions, Atlassian generates a huge amount of excess
cash by the end of the 5-year projection period. How do you think it should use this
cash? As an investor, how might you take advantage of this excess cash balance to
mitigate risk?
The best use of cash seems to be increased investment in sales & marketing, or perhaps CapEx
used to fund overseas expansion into emerging markets.
It is unclear how productive additional sales & marketing spending would be weve come up
with our own estimates for customer acquisition costs, but does each dollar spent actually
result in progress toward a new customer?
It is also unknown what impact commissioned sales reps would have it may be that until the
company starts expanding its product offerings and going after more mainstream users,
additional sales reps may not make much of a difference.
CapEx requirements are likely to increase as more and more customers switch to the
OnDemand versions of the companys products, so that could be a more sensible way to
spend money. Overseas expansion, especially in relatively underpenetrated emerging markets,
would also constitute a good use of CapEx.
Given that Atlassian generates a large excess cash balance by the end of Year 5 in our baseline
scenario, we could argue that the company should issue a dividend to equity investors (Accel
and T. Rowe Price).
This would allow them to realize some of their returns earlier on; the problem, however, is that
the two firms combined own less than 20% of the company.
So such a strategy may not even make much of a difference. It would be more viable as a way
to mitigate risk if the company had a true cash cow business with much higher margins and
an even higher excess cash balance, and/or if the equity investors owned a much higher
percentage of the company.

04 Atlassian Case Study Discussion PDF

Uploaded by

Copyright:

Available Formats

04 Atlassian Case Study Discussion PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

04 Atlassian Case Study Discussion PDF

Uploaded by

Copyright:

Available Formats

Atlassian Case Study Answers and Discussion

The companys customer base has to triple;

Of course, there are other ways to get to the same outcome.

It is actually targeting a specific IRR rather than a money-on-money multiple.

If it really is profitable, it is likely because it is spending less on sales & marketing as a

Threat of Substitute Products/Services: There is a high threat of substitute

Bargaining Power of Customers (Buyers): In most of these markets, there is substantial

Intensity of Competitive Rivalry: As mentioned above, these markets (online file

Atlassians market, in particular, has dozens of competitors offering slightly different

You might also like