Cimpress (CMPR)
Guest Write-up by Jayden Preston
Introduction
Vistaprint was founded in 1994 by Robert Keane, on the idea that there was a market for producing business cards cheaply for companies too small to order in large quantities. Through utilizing the internet, creating better software and printing technologies to make such mass customization cost effective and efficient, the business has turned out to be a huge success.
23 years later, Vistaprint has become a clear leader in mass customization of marketing materials for small businesses. It has also increased its portfolio of businesses and target markets through acquisitions and now offers thousands of products that customers can customize to their needs. It was renamed to Cimpress in 2014 to reflect their growing number of business units.
Business Summaries
Cimpress has changed their business segmentation a number of times in the past. We look at their current segmentation below:
1. Vistaprint
This is still their core business. Vistaprint currently operates globally, with a special focus on North America, Europe, Australia and New Zealand markets. Their Webs-branded business is also grouped under Vistaprint.
The target customers of Vistaprint are micro businesses, i.e. companies with fewer than 10 employees. Since they are small operations without marketing expertise, they rely heavily on the software technologies provided by Vistaprint when designing and making their marketing materials.
In FY2017, Vistaprint served 17 million micro businesses and generated $1.3 billion in revenue, representing 60% of Cimpress’s overall revenue. Adjusted net segment operating profits were $165 million, for a margin of 12.6%. This is a depressed margin though. The two-year average from FY2015 and FY2016 is higher than 17%.
2. Upload and Print
This segment is mostly built up through acquisitions. Companies/brands in this segment include: druck.at, Easyflyer, Exagroup, Pixartprinting, Printdeal, Tradeprint, and WIRmachenDRUCK businesses. These companies mostly focus on Europe.
Some companies that have more resources would look for professional graphic designers, ranging from local printers, print resellers and graphic artists to advertising agencies and other customers with professional desktop publishing skill sets, for their marketing needs. The segment focuses on serving such graphic professionals.
As a group, they generated $589 million in revenue in 2016, which is 28% of the total. Adjusted net segment operating profits were $64 million, for a margin of 10.9%
3. National Pen
This business was acquired at the end of 2016. Given its scale and vertical integration, National Pen is the leading provider of a wide array of customized writing instruments for small- and medium-sized businesses in more than 20 countries. The company also sells other promotional products, including travel mugs, water bottles, tech gadgets and trade show items.
It serves about a million small business customers through a successful direct mail marketing and telesales approach, as well as a small growing e-commerce business.
The business generates around $113 million in revenue.
4. All Other Businesses
This segment includes the recently sold Albumprinter, Most of World, and Corporate Solutions businesses. These businesses have been combined into one reportable segment based on materiality.
This segment contributed $129 million in revenue but had an adjusted operating loss.
Moat vs Durability and Quality
Cimpress is a fascinating company to study and contemplate. It’s a controversial stock, even though it might not appear so at first. While the shares are, or were owned, by value investing legends like Allan Mecham and Glenn Greenberg, there are also significant short interests, like 20% of the float, in the stock for a long time. The crux of this article is to explain why this is the case.
Despite the name change, Vistaprint is still the key driver of value creation in Cimpress. As an upshot, we will focus our discussion on Vistaprint. But I might refer to Cimpress when I am quoting numbers for the whole company. In principle, this should not affect the message.
Let’s start with moat this time.
Moat
Vistaprint is the poster child of a scale business.
When asked about what type of moat he favors most, Prof. Sanjay Bakshi, the famous value investor and professor in India, picked cost advantages. Buffett has expressed similar preferences if my memory serves me right. And within sources of cost advantages, scale is clearly the most attractive. Scale creates the same sort of fly wheel effect popularized by Bezos and his lieutenants at Amazon. In short, it is a virtuous cycle where cheap prices feed on volume growth; volume growth leads to lower cost structure, which then renders cheaper prices possible.
The same goes for Vistaprint. Every increment of growth in volume increases their competitive advantage. The longer you keep a printing machine working, the lower the production cost will be as you are spreading your fixed cost over a larger amount of products. This is the reason why historically, you can only get a low quote if you order in bulk. And to do bulk production, you had to offer more standard designs. Vistaprint solved this issue by their software and sheer amount of scale not achieved by any local offline printers. Cimpress, as a whole, now produces more than 46 million unique ordered items a year. That’s more than 126,000 items per day. Importantly, this allows them to aggregate individual orders and reach scale at the individual design level, making mass customization economical. Clearly, no competitor receives as many daily orders as Cimpress, allowing them to become the industry’s lowest cost provider.
Business cards are the bread and butter of Vistaprint’s business, and even now business cards still account for 30% of total bookings. And it is used as the paragon for their scale advantage. The following is what they say in their latest 10K:
“The current Vistaprint operations for a typical order of 250 business cards in Europe and North America require less than 14 seconds of human labor for all of pre-press, printing, cutting and packaging, versus an hour or more for traditional printers. Combined with advantages of scale in purchasing of materials, our self-service ordering, pre-press automation, auto-scheduling and automated manufacturing processes enable us to produce standard business cards at a fraction of the cost of typical traditional printers with very consistent quality and delivery reliability. Achieving this type of efficiency took us more than a decade and required massive volume, significant engineering investments and significant capital.”
Extending their discussion on their scale advantages to other products:
“We believe that the potential for scale-based advantages are not limited to large, automated production lines. Other advantages include the ability to systematically and automatically sort through the voluminous “long tail” of diverse and uncommon orders in order to group them into more homogeneous categories, and to route them to production nodes that are specialized for that category of operations and/or which are geographically proximate to the customer. In such cases, even though the daily production volume of a given production node is small in comparison to our highest volume production lines, the homogeneity and volume we are able to achieve is nonetheless significant relative to traditional suppliers of the long tail product in question; thus our relative efficiency gains remain substantial.”
The scale advantage doesn’t only exist in lower production costs. Last year, Cimpress spent $243 million on technology and development expense, including $52 million spent on R&D. These are huge numbers relative to their competitors. For instance, according to Mecham, the industry average technology and development spending was $11 million in 2013. Back then Vistaprint was spending $164 million already. That’s 15 times as much as the industry average. We can see that Vistaprint is now spending even more relative to its competitors, likely further increasing their advantage in technologies.
Cimpress also spent $611 million on marketing and selling expense in FY2017, including $364 million alone on advertising. This creates a similar edge with the cost of acquiring new customers. A large amount of small businesses still buy at the local, mom-and-pop print shop – a source that is in terminal decline due to the negative feedback loop they face. (Lowering revenue leads to less resources for investments, rendering their services less attractive and so on) Since 2004, over one-third of commercial printers have gone out of business. None of these small competitors can afford any national advertising campaigns. As Cimpress grows, the mindshare of their brands should only grow.
All in all, these three pillars together create a formidable feedback loop: more customers lead to greater scale, which allows better technology, a wider range of product offerings with better services and lower pricing, which then lead to more customers.
The printing industry will continue to exist. Nevertheless, it will be a more consolidated market in the future with a lot less local small printers. On the other hand, it looks clear that Cimpress will be a dominant “survivor”. In fact, the runway for growth seems immense for them. The company now has revenue of $2 billion. Yet, they are estimating revenue opportunity for low-to-medium order quantities to be over $100 billion annually in North America and Europe alone.
One usual concern on the moat of Cimpress is whether it will stay intact under a potential onslaught by a huge scale player like Amazon. Admittedly, given Amazon’s nature of extreme willingness to endure losses, in theory they can spend hundreds of millions during, let’s say, a 5-year period to build up their technology and printing systems, as well as their brand. In practice though, as Amazon is going for markets, such as grocery, that are counted in the trillions, the whole mass customization market just doesn’t seem as attractive for them. More importantly, as the market is still shifting from offline to online, the major competition ground is still between online and offline players. Amazon’s presence might foster this transition.
Management and Capital Allocation
The case for Cimpress is usually further bolstered by the management team. Founder Robert Keane currently owns around 10% of Cimpress. He is also willing to admit mistakes and learn from them. His shareholder letters are some of the most unique documents you will read from a CEO. In them he details the way they think about the business, the capital allocation policies they are pursuing, the mistake they have made and how they plan to move forward. With lessons learned from past capital allocation mistakes, the company now has more specific hurdle rates for investments with varying risks. And it’s not only acquisitions they are considering. They are now more cognizant of share buybacks as a potential opportunity cost when buying back shares offers the best returns. Last but not least, Keane is willing to endure short term pain for benefits down the road, as demonstrated by the significant investments they have made. The Company then details all the organic investments and gives shareholders a range of estimates of investments for growth to arrive at what they call the “steady state” free cash flow.
Moat Is All You Need?
Value investors, under the influence of Buffett and Munger, love the concept of moat and competitive advantages. To a large extent, when they research a company, moat is all they focus on. Moreover, ever since popular books such as “Competition Demystified” came out, some investors take it a step further and concentrate only on barriers to entry.
This is understandable given the importance of moat in fending off competition in a capitalist world. In the case of Cimpress, with a moat that is expected to be durable and a big market ahead, along with a management team that treats capital allocation seriously, this makes for a great recipe for most value investors. It seems all you need to do is to buy this company at an attractive valuation.
Allan Mecham said as much himself back in 2013/14:
“Key to any good business analysis is identifying levers of competitive advantage (i.e., the moat) and figuring out if they’re durable. In VPRT’s case, the moat appears deep and durable and is based upon scale. As VPRT grows, unit costs decline…
“We take comfort being co-owners with VPRT’s management team. Keane and company are large shareholders and have demonstrated savvy capital allocation decisions, balancing the long-term investment needs of the business with outside opportunities, such as acquisitions and share buybacks. VPRT is a nice addition to the portfolio and has a long runway ahead of it.”
Nonetheless, this is where I believe most investors are incomplete with their investment or business analysis. While moat is important, it is not sufficient for a business to be a good one. It’s crucial that investors not equate the two concepts.
Let me explain.
Quality and Durability -> Product Economics
Many times this is obvious. In terms of barriers to entry, telecom businesses should rank among the top. The biggest telecom providers usually command a substantial scale advantage that no new comers can hope to achieve organically. As they have more subscribers, more resources will be at their disposal to subsidize smartphones, invest in data transfer technology and whatnot. It’s also exorbitant for a new entrant to replicate the network of an existing giant. Their barriers to entry are strong. However, few value investors view them as high quality businesses. The problem for the telecom industry is that for them to deliver their services to the customers, a huge amount of capital expenditure needs to be made. This is compounded by the need to revamp the whole network as the next generation of data transfer technology arrives. PwC* once estimated the industry average ROIC to be below their cost of capital. In this case, while moat exists, it’s clear that the product economics for a telecom business is bad.
While moat analysis works on a “macro level”, product economics concerns the “micro”. Only when there exists an attractive “profits to capital invested ratio” at the product level can the business be of high quality.
This above message sounds needlessly simple. After all, what I am really saying is that moat is only meaningful when the economic castle in the middle is valuable. And the metric to assess its value is return on capital. This is widely understood.
The key issue is investors seldom distinguish between a castle that is of value now and a castle that will stay valuable in the foreseeable future. We are taking about “Durability” here. Kodak’s economic castle was very strong. Their mind share was immense and they had a very profitable business. Yet it no longer exists because the camera film they sell is no longer desirable. The lack of durability of camera film showed Kodak to its grave. (We can further attribute their demise to management’s inability to accept new technologies, even one from their own creation. But the point remains that their key product wasn’t durable).
Herein lies the issue Geoff and I see in Cimpress despite its historically attractive return on tangible capital: The high churn within its customer base. We view this as a major weakness in their product economics.
Partly because of the nature of micro businesses (~20% of small businesses go out of business, with ~20% being started each year) and partly because of the products they sell (discretionary products for small businesses such as business cards), their retention ratio (year over year) has historically hovered around 20% to 30%. This means to retain the same earning power, Vistaprint needs to find enough new customers to fill up 70% or more of their existing customer base next year. While they have handily beaten this hurdle in the past, this challenge will only compound itself should they fail to increase their retention in a meaningful way in the future. Management has said their target is for Vistaprint to increase its revenue by double digits in the near future. Assuming they do so through an augmenting customer base, i.e. Life Time Value of customer staying flat, you can calculate that for the company to grow 11.3% annually for 10 years, the rate of increase in the number of new customer needs to be 14% p.a., reaching 31.5 million. (Here I assume the base metrics to be from 2014. That was the last time they disclosed the numbers of new customers vs retained customers. And here they define retained customer as any unique customer who has also purchased in any prior period, i.e. not a year over year figure. I also erroneously assume the retention rate remains the same throughout the lifecycle of a customer. Please just take these numbers as very rough estimates/indications).
To put these numbers into context, notice that there are approximately 60 million businesses with fewer than 10 employees (Vistaprint’s main target customers) in the US, Canada and European Union. If Vistaprint can continue to grow at double digit rates to 2024, they would need to acquire more than 50% of the total market in a year by then. This simply cannot be sustained. Essentially, to grow the business, they had relied a lot on “feeding their acquisition machine”. To sustain growth, they need to improve retention rate, average booking, gross margin, or a combination of the above.
This reminds me of multi-level marketing companies, sans the ethical issues. Most MLMs have high returns on capital because they have low costs in distributing and marketing their products. And they can stay very profitable for a long time. However, the high churn within their distributor base creates a problem that eventually needs to be dealt with: There will be a time when you run out of people to scam.
Management is not ignorant of this problem, which they refer to as the “Leaky Bathtub”. To make growth sustainable, they started working on fixing the leakage. Since 2012, they have offered fewer discounts, lowered shipping fees, been less aggressive in cross-selling, and invested into expanding their product portfolio and lowering promotionality to create a more trusting relationship with customers. Basically, they are trying to find customers whom they can build longer term relationships with.
The problem is without focusing on the low customer retention, it’s difficult to foresee such problems before they arrive. Had you been looking at their financials prior to 2012, you would believe the company to be in good shape. From FY2001 to FY2011, the Company grew its revenue by 63% annually. The annualized growth was still strong at 40% from FY2006 to FY2011. In fact, even management was late to identify the core underlying problem. At the 2012 Investor Day, management highlighted their expectation of further gross margin expansion from the industry leading level of 65%. Through a combination of bookings growth (low to high single digit growth), number of customers increase (in the teens for new customer) and higher mix of retained customers, they believed annual growth rate in revenue could be sustained at 20%. Back then, their estimate for Life Time Value of new customers was $145, with the assumption that gross margin will stay at 65%. This is super attractive when compared to the $27 in estimated average customer acquisition costs. As they increase their gross margin and acquire higher value customers, they believed advertising cost as a percentage of revenue could eventually go down by 2017.
Instead, retention rate (year over year) decreased as they repositioned the brand away from price-oriented customers, reaching a new low of 26% in FY2014, down from 30% in 2011. Number of customers grew only 3% p.a. from FY2011 to FY2017 within Vistaprint. As a result, revenue only grew 8% a year for Vistaprint since 2011 to now, including acquisitions. Their adjusted ROIC also dropped from 30%+ to the mid-teens. At the same time, instead of expanding, gross margin went down to 61% due to all the changes they have implemented in the business. Now, it’s difficult to estimate how Life Time Value has changed from 2012 because they stopped disclosing detailed operating metrics. By way of example, we no longer know how many new customers the business is acquiring each year. If we use change in average revenue per unique customer as a proxy, the increase is likely to be only 1.3% a year from 2011. Worse still, gross margin has decreased to 61% from 65%, or a decline of 1.1% p.a. The good news is all the investments made to transform Vistaprint since 2012 are seemingly working. After dropping to a low in FY2014, retention rate last fiscal year went up to 31%, comparable to the level in FY2011. Notwithstanding this positive development, I do not believe Life Time Value of new customers for Vistaprint has increased much, if at all, over the last 5 to 6 years. After all, they now have the same customer retention figure 6 years ago, but with a lower gross margin.
Advertising cost as a percentage of revenue for Cimpress did decline from 22% in FY2012 to 17% in FY2017. Yet, this is mainly due to much higher revenue per customer within the new Upload and Print Segment, where marketing, service and technology cost in total is just 14% of revenue. For Vistaprint alone, I do not believe advertising cost as a percentage of revenue has gone lower. In fact, it’s easier to believe it’s gone up given all the changes they have made.
The Uncertain Value of Growth
Altogether, this points to the difficulty in valuing growth for such a company. Were the investments spent prior to 2011 valuable in the sense that they increased the earnings power of the company more or less permanently? How about the investments between 2012 and 2017? Retention rate has now recovered, but could we be sure product economics will improve by 2022? Management gives out their estimates for maintenance investments. Yet as recent business performance has demonstrated, it’s not easy to delineate maintenance investments from true growth investments.
In fact, when we look at the track record between 2011 to now, we have good reasons to doubt the value of all the investments. The 3-year average EBITDA per share in 2011, i.e. average from 2009 to 2011, was $2.70. The figure for 2017 was $5.50. That’s good for an annualized growth of close to 13%. However, it’s misleading if left out of the context of how they funded the growth. Back in 2011, they had net cash of $5.30 per share. Lately, that has turned into a net debt position of close to $1.1 billion, or around $34 per share (Here I include capital lease obligations and earn-out liabilities). As an upshot, the more than $2.8 a share increase in EBITDA was achieved by leveraging up the balance sheet, in a substantial way. If we think in terms of “the price” they have paid for the increase in EBITDA, it’s 13.8x EBITDA. I will not encourage anyone to buy assets at this valuation.
One could argue that the past few years has been tough on them as they underwent a period of transition. Yet this is exactly the point: It’s difficult to value growth initiatives before the facts crystalize, since there is little replicability in the business. Without high retention rates, high replicability is difficult to achieve since you are mostly dealing with new customers every year.
Does this mean we should only invest in businesses with high retention rates? Geoff is very wary of the product economics in Vistaprint, yet he invested in another business with low retention rate, Weight Watchers. He wasn’t contradicting himself.
Both Cimpress and Weight Watchers are the dominant player within their business domains, enjoying attractive return on invested capital. The major difference between them, however, is management’s attitude toward growth. While Weight Watchers had been busy buying back shares, Cimpress has always had more of a growth oriented mindset. From FY2011 to FY2017, they spent more than $1.3 billion on capital expenditure and acquisitions. (Some of the capex should be regarded as maintenance capex). As mentioned, there were also non-capitalized expenses that they regard as organic growth investments. Those together could easily add up to much more than $400 million. In fact, the high end of the estimated amount of organic investments not needed for steady state cash flow given by management is more than $394 million for the past 3 years alone. Let’s randomly assert maintenance capex roughly equals non-capitalized expenses for growth, so they offset each other. Just looking at the $1.3 billion figure, that’s nearly as much as the market capitalization by the end of 2011. Had they not done any investments, they could perhaps have bought back half of their shares outstanding while staying debt free now.
Valuation
Let’s still take a crack at valuing the business, despite the obvious difficulty. Please bear in mind that all these are rough estimations.
Method 1
We will start off with the approach management outlined. Yet, I do not recommend this “academic” try at valuation.
In the shareholder letters, Keane explained how they measure their intrinsic value per share:
We estimate the pre-debt value of a single share of Cimpress to be equal to:
(i) unlevered steady state free cash flow per share divided by our WACC;
plus
(ii) the per share net present value of future expected returns on the capital not required to maintain our steady state that we have allocated in the past and may allocate in the future.
I see a problem with this valuation method. As their steady state free cash flow includes organic investments for growth, if they then add the net present value of future expected returns on organic investments for growth, I see the potential for double counting the value of organic investments for growth in this equation.
It makes more sense to me if in part (i), we simply divide the adjusted actual free cash flow by their estimated WACC. Here I mean free cash flow, after deducting all the organic investments. Then we add the value of organic investments for growth.
From FY2015 to FY2017, the average adjusted actual free cash flow was $124 million. This is an unleveraged figure. Dividing that number by their WACC of 8.5%, gives us a valuation for that stream of cash flow of $1.463 billion. Our estimate for yearly organic investments for growth comes from the two-year average of FY2016 and FY2017. (These are the two years where management gave us their estimates). The high-end number is $224 million, while the low end is $159 million. One key assumption we need to make here is the after tax return they can get from such organic investments. We will use 10% for the low-end figure and 15% for the high-end figure to assemble a range of valuation for Cimpress. Next, we calculate what kind of NPV each year’s investments should be. If $1 dollar can generate 10% a year, with a cost of capital at 8.5%, the $1-dollar investment should be worth $1.12. If you can generate 15%, then your investment should be worth $1.59. We can then add a stream of such yearly investment’s NPVs, properly discounted by the WACC, to arrive at the value of all the organic investments for growth. (We are thus assuming they can deploy as much capital with the same returns every year going forward). I calculated the range to be $2.278 billion to $4.534 billion. The company has around $1 billion in net debt (before the proceeds to be received from their sales of Albumprinter and including capital leases). This gives us a range of estimated equity value of Cimpress at $2.741 billion to $4.997 billion.
The upper bound seems unreasonable to me. Not only do you need to fully trust the high-end estimate of growth investments by management, you also need to believe they can make 15% year in and year out for a long period of time. Besides, there will be investments where they have a lower hurdle rate due to more certainty. Even with this “best case” scenario, you could make 50% at most on the stock at the current price level.
Method 2
The more sensible approach would be to estimate their free cash flow 10 years out and then figure out the likely valuation the company could get at that level of free cash flow.
Again, we use the free cash flow excluding all organic investments. However, we are not using what management showed us. Instead, we will use the leveraged free cash flow figure. This entails deducting cash outflow from withholding taxes in connection with equity awards and payments toward capital lease obligations, both buried in the “Financing Activities” section of the cash flow statement and not highlighted anywhere by management. From FY2015 to FY2017, these two items would have deducted $29 million a year on average from free cash flow. This is not insignificant at all. Calculating free cash flow the traditional way, i.e. cash flow from operation minus capex and capitalized cost for software development cost, alongside the two outflows stated above, this brings us a 3-year average free cash flow estimate of $85 million from FY2015 to FY2017.
I will assume a more a reasonable return on investment figure of 12%. Throughout 10 years, they would then have invested $1,590 million to $2,240 million for growth, and generating an extra $190.8 million to $268.8 million a year in free cash flow. ($159 million is their low-end estimates for organic investments for growth and $224 million for the high end). As a corollary, by 2027, they could be generating $276 million to $354 million in free cash flow.
If the company can command a multiple of 30x free cash flow, the equity value of Cimpress could be $8.3 billion to $10.6 billion by 2027. Buying the stock now would get you an annual return of around 9% to 12%.
Method 3
Last but not least, let’s do a final valuation with the simplest methodology of the bunch here.
10 year average EBITDA margin is 13.1% for Cimpress. Applying that to the latest revenue gives us a normalized EBITDA estimates of $278.8 million. At a better than average valuation of 10x EBITDA, enterprise level should be $2.8 billion. The equity value alone exceeds this level now. In fact, the current enterprise value of Cimpress is at $4.4 billion, giving the company an EV/EBITDA multiple of 15.7x
The only way to believe Cimpress is worth more than 15x EV/EBITDA is to believe the investments above the EBITDA line are all generating attractive returns. Only so could EBITDA be growing at a rate that justifies the current EV/EBITDA multiple.
Conclusion
It’s never easy to understand market behavior. But let’s ask again why there is a 20% short interest in Cimpress. Worrying about the durability of physical business cards or other major physical forms of advertising could be one reason. Fearing they would be toast when Amazon comes in with full force may be a better one. In my opinion, the best reason to have reservations about Cimpress is its poor product economics in addition to management’s growth aggression.
Small businesses don’t always have the need to get business cards or other marketing materials. Such discretionary spending leads to very low buying frequency. Most Vistaprint customers buy only once or twice a year. And each time, the orders are only around $35. Worst yet, close to 70% of them don’t purchase from Vistaprint again next year. Operating in the small business segment, Cimpress has an inherent difficulty in improving the retention rate**.
With such product economics, it’s simply too difficult to have confidence in aggressive growth initiatives. And without getting confident about the return they can receive from their organic investments, there is no way to see Cimpress as an attractive stock now.
Should their Upload and Print segment, where better product economics exist, grow to be a much larger part of Cimpress, or the management stops being so aggressive in growing the business, Cimpress could become a very attractive investment candidate. But not now, despite its durable moat.
* https://www.pwc.com/id/en/publications/assets/ticepublications/pwc-capex-playbook-telecoms.pdf
** Management’s comment on retention rate on 2012 Investor Day: So we’re not aiming to, for instance, get that retention metric from 42% to 80%. That would not be realistic, given the nature of the small business population. (Page 27 of the 2012 Investor Day Transcript)