Description

What would you pay for an industrials business with 60% gross margins and 25-30% EBIT margins that grows revenue like this?

And then what about one that grows like this, only now with 800-900bps lower gross margins and 1000bps lower EBIT margins, emerging competition, warning signs around its TAM, and a Chief Revenue Officer who quit barely more than a year after joining? 

If you said “basically the same multiple”, then you are probably long Protolabs (PRLB). We think PRLB used to be a great business. We’d still call it a good business today. But we also think it’s a victim of its own success and the stock is a short at current levels (near $107, 38x 2019 / 33x 2020 consensus EPS) with 40-50% potential downside..

*PRLB was written up as a short by tac007 in May 2017 which you can read for added context

Situation Overview

PRLB has been an investor growth darling for much of the time since its 2012 IPO, trading at a premium multiple that reflected its perception as something akin to an industrial unicorn – a high-growth, high-margin, seemingly acyclical parts manufacturer. 2016 revealed some cracks in the bull story (full-year organic growth decelerated every quarter and Q416 ended up negative) but the stock rebounded to reach near peak valuation levels in 2017/2018 as growth recovered (helped in part by M&A) and tax reform turbocharged earnings. While management and Street analysts point to various reasons why 2016 should be excused as an isolated hiccup, we believe it likely signaled a structural shift in the business – a new-normal of more competition, lower growth, lower margins, and more cyclicality as the company tries to tilt its business more towards production vs. prototyping. 

2019 will end up being another bad year by PRLB’s standards (~2% revenue growth, the lowest level since the company went public in 2012; -10% EPS decline, the 2nd time in the last four years we have seen negative earnings growth) … estimates have come down yet the stock is basically back to the same levels it traded following their Q4 report a year ago. Hard data for industrials/manufacturing remain soft, but depending on which macroeconomist you listen to, LEIs for economic expansion and industrial activity suggest a recovery is coming - it certainly seems like PRLB investors have been willing to bet on an imminent return to the days of 15%+ EPS growth ... we are skeptical given the changes to the business we have observed.

Business Overview

Protolabs is an industry leader in quick-turn manufacturing, a service primarily used in prototyping applications and short production runs. The company offers injection molding, CNC machining, 3D printing, and sheet metal capabilities to a customer base of ~45k+ unique engineers (its largest customer is ~1-2% of sales - we believe this may be Tesla). Founded in 1999, PRLB was an early innovator in software algorithms to enable automated manufacturing. This was accomplished by defining a relatively narrow scope (“box”) of materials, sizes, and geometric complexities for parts that PRLB would be willing to manufacture. Engineers could upload CAD files to the PRLB website and the company’s software algorithms would then analyze the part design to see if it fit the “box”, generate an automated price quote if it did (or manufacturability design change suggestions if not), and transmit approved orders to one of the machines in its manufacturing facilities to produce with minimal human intervention required, This enabled PRLB to fill orders in days instead of weeks/months and at competitive prices while still earning 60% gross margins, 25-30% EBIT margins, and 20%+ ROIC/ROEs historically. That financial profile has been diluted down over time as the 3D printing and sheet metal businesses they’ve acquired in recent years are lower-return businesses but enabled PRLB to offer a more expansive set of solutions to address customer needs.

Roughly ~80% of sales are to US customers, ~18% to Europe, with the rest to Japan. Its end-market mix is spread across auto, aerospace, healthcare/med devices, and general manufacturing.

Key Thesis Points

1)    2016 was not an isolated hiccup but rather a signpost of a shift to a new-normal growth and margin profile for PRLB

•  We believe this new normal is primarily the result of two factors: (1) a shrinking white-space opportunity in the quick-turn Injection Molding and CNC machining businesses (evidenced by a long-term slowing trend in organic growth and its unique developer count KPI, adjacent M&A into 3D printing and sheet metal, discontinuation of certain service offerings due to lack of demand, and changes in the sales model to tilt more towards “enterprise” customers and short-run production); (2) PRLB’s moat and pricing power being eroded by new competition. In our view, the days of acylical growth are in the rearview mirror and PRLB’s future will be more tied to the normal business cycle ups & downs of traditional manufacturers

• Prior to 2014, PRLB was able to achieve 25%-30%+ organic revenue growth regardless of the macroeconomic backdrop. Its innovation filled the gap between local machine shops (willing to do prototyping and short production runs, but often with 4-8+ week turnaround times due to the significant amount of human intervention required) and much larger contract manufacturers (designed for mass production, not one-off prototyping jobs or small part runs), resulting in a market share opportunity that more than offset any macro cyclicality

• Starting in 2014, however, a multi-quarter trend of slowing “core” (defined as the legacy injection molding and CNC machining business) revenue growth and margin degradation coincided with the emergence of new competition and strategy shifts referenced above. This slowdown was somewhat masked by M&A (Fineline in 2014), which was a signpost for future slowdowns in the core as well (Alphaform in 2015, RAPID in 2018)…

•  We believe increased competition is a key reason why PRLB has been unable to sustain prior levels of growth. Several well-funded “copy-cat” companies have emerged since PRLB’s IPO in 2012 (Xometry, ZYCI, Plethora, to name a few) and several existing competitors (Xcentric, ICOMold) adapted their offerings to compete more effectively with PRLB. Whereas PRLB historically had the quick-turn manufacturing niche largely to itself, competition seems to be crowding out some of the white-space. 

https://www.prnewswire.com/news-releases/xometry-announces-50mm-equity-raise-led-by-greenspring-associates-300848569.html

https://www.engineering.com/AdvancedManufacturing/ArticleID/16449/Why-On-Demand-Could-Be-the-Future-of-Manufacturing.aspx

  i.      On the software side, competitors have already replicated many of the key features in PRLB’s offering (CAD file uploads for instant price quoting, faster turnaround times, variable pricing based on delivery speed). While no one has been able to fully replicate the manufacturability analysis that PRLB provides as far as we can tell, competitors have approached this from a different angle than PRLB. Rather than trying to fully replicate its automation, competitors have offered a hybrid approach – more automation than a local shop but more customization options than PRLB will provide. This results in a lower GM% and longer average turnaround times, but greatly expands the number of developer needs these companies can address. Smaller customers (still the majority of PRLB’s revenue base) now have multiple alternatives to decide how to trade-off price vs. turnaround times vs. customization needs. And larger customers can now consolidate their outsourced prototyping with a single, more flexible vendor rather than spread projects across multiple vendors, each with their own defined “box” limitation

• Another sign of competition seems to be reflected in a now multi-year trend of gross margin degradation. Even if we adjust for the dilutive margin impacts of the 3D printing (Fineline, Alphaform) and sheet metal (RAPID) acquisitions, our math suggests gross margins in the legacy core (Protomold and Firstcut) appear to be in multi-year decline, approximating several hundred basis points.

2)  Bulls are anchored to past performance and are underestimating the economic sensitivity of the current business

• Many bulls point to flattish revenue growth in 2009 as a sign of how resilient the business is but this data point is misleading in three key ways: (1) ProtoMold (the injection molding division) revenues were actually down -8% that year; it was only the launch of FirstCut (the CNC machining business) that helped mitigate those declines on a total reported revenue basis; (2) PRLB was significantly smaller at the time and still in the earlier innings of expanding its product offering ($40mm of revenues then vs. $450mm+ today); and (3) PRLB has since tilted its business increasingly towards production (more economically sensitive and cyclical than prototyping). Further, given we have now observed two significant slowdowns in just the last four years, we find it hard to believe that some analysts still defend PRLB as being less cyclically exposed

• As it relates to “upside” cyclicality, it is worth noting that PRLB’s business does not really benefit from traditional scale economies – in fact, there is asymmetric operating leverage (& earnings risk) to the downside. This asymmetry is due to the way PRLB must utilize capacity in order to run its quick-turn model. Management has described peak utilization as being capped around 80% - anything higher than that and it risks not being able to meet orders for its most differentiated and most profitable business offering - 1-2 day turnaround “rush” deliveries (that most competitors can’t match yet). To avoid that, management must continually invest in new plant & equipment capacity as revenue grows in order to meet its turnaround time promises, which in turn makes it difficult to ever achieve significant operating leverage on its existing fixed cost base

• As simple evidence of this, even if we ignore the issues of 2016 and just look at 2010-2015, revenues grew 3x over that period ($50mm to $162mm) yet EBIT margins actually declined 10bps (25.7% to 25.6%)

• In downturns, however, PRLB retains the deleveraging risk of a traditional manufacturer. To make matters worse, fixed cost absorption issues are typically magnified by pricing pressures, as the fragmented base of competing local job shops cut price to bring in volumes and can offer shorter turnaround times given there are less orders to deal with. While there is not extensive precedence for this given PRLB’s high-growth history, we can look to 2009 and 2016 as potential warnings signs of what can happen in a slowdown

• In 2009, revenues declined 1.5% but EBIT margins deleveraged 600bps on a 240bps drop in gross margins. 2016 has some noise from a dilutive acquisition made in early Q415, but if we look at just the Q416 quarter, revenues declined 2% y/y but EBIT margins declined 210bps (and we estimate would be closer to 300bps on a comparable basis

3)   PRLB’s entry into 3D printing, while helping mask the slowdown in legacy product growth, risks cannibalizing that same core while also dragging down margins & returns going forward

•  While 3D printing is used extensively in prototyping applications (and increasingly in parts production), in the years prior to its acquisitions of Fineline (2014) and Alphaform (2015), PRLB’s senior management had characterized it as being a less differentiated, more competitive market and not a real threat to its core business. As such, it was something of a surprise when they subsequently decided to acquire two 3D printing service bureaus. Management’s story was that the 3D printing assets they acquired fit into an earlier stage of the prototyping process and they wanted to offer an end-to-end solution through the prototype iteration lifecycle to customers (unlike DDD/SSYS/HP, which also offer services, PRLB does not manufacture their own printers or control the underlying technology)

•  At the same, while it may just be a coincidence, the first deal (Fineline) coincided with nearly the exact quarter when core growth started to slow (Q2’14) and the second deal (Alphaform) came shortly before it dropped off a cliff (Q4’15). And while those deals helped to mask the slowdown, this came at the expense of margins and business quality. This is because there are lower barriers to entry (anyone can buy the same 3D printers that PRLB uses), more competition, and less service differentiation, which manifests itself in lower margins (earnings call & SEC filing disclosures let you estimate that the blended gross margins on the acquired businesses are ~40% vs. PRLB’s existing core of ~55-60%)

•  At the same time, continued advances in 3D printing technology have allowed it to increasingly blur the quality line between a 3D printed part and an injection molded/CNC machined part, increasing substitution risk. PRLB acknowledged as much in its Q3/Q4’16 calls, suggesting that part of the reason it was seeing slower core growth relates to substitution of 3D printing services for CNC machined plastic parts in certain instances

• PRLB’s late 2017 acquisition of RAPID appears to have similar trade-offs. While sheet metal has less substitution risk, it is still a lower-margin, lower return business than PRLB’s legacy businesses. As referenced earlier, one spin would be to say that it complements PRLB’s existing offerings and helps it capture a higher share of wallet, making its customers stickier in a way that simple math may not reveal. But the skeptics argument would be that it represents another example of “de-worsification” and a signal that management sees a shrinking white-space in that legacy core

Key Risks

1. Organic growth meaningfully reaccelerates (either due to cyclicality or because we are underestimating the amount of white-space opportunity still remaining)

2. Accretive M&A (with the possible exception of Fineline, the acquisitions done to date have been value destructive and, specifically to RAPID, created organizational headaches that are still not solved … we think this makes PRLB increasingly hesitant to buy their way out of growth slowdowns again)

3. More aggressive balance sheet utilization (there is a ~$50mm buyback program in place = ~2% of market cap and the company has an unlevered balance sheet)

4. Investor psychology (more specifically, the bulk of our returns on this short will likely result from a multiple rerating if/when other investors acknowledge the structural changes we have identified ...  that is something we can’t predict when or how it will occur)

Valuation

Bulls point to PRLB’s historical average multiple of 30x forward earnings to suggest there isn’t much more downside rerating risk, but we think they are overlooking the fact that future growth is likely to be structurally lower and more cyclical than it has been in the past, and the theoretical white-space opportunity is shrinking, both of which suggest it deserves a multiple reset.

Looking at a peer set of other “high-quality” industrial businesses (ex. TDG, DHR, APH, ROP, FAST, etc), we struggle to see why PRLB should trade at a significant premium and we would argue it should trade at a discount. At 20x forward earnings (which we think is still relatively generous but also where it traded in 2016 during the bottom of its prior slowdown), we model ~40-50% downside potential to the stock.

Catalyst

Continuation of below-average growth results in limited/negative earnings growth + multiple compression. PRLB reports Q4 results on 2/6 and typically provides 1 quarter forward guidance (there is typically just ~2weeks forward visibility in the business model given the short average lead-times for most projects).