“One metric catches people. We prefer businesses that drown in cash. An example of a different business is construction equipment. You work hard all year and there is your profit sitting in the yard. We avoid businesses like that. We prefer those that can write us a check at the end of the year.”
Charles Munger, 2008 Berkshire Hathaway AGM
I was introduced to Pulse Seismic (PSD.TO) by Mackie who authors the excellent Moatology blog. Pulse is a Canadian microcap. Its business should be the easiest in the world to understand. They own a library of seismic data (“seismic data” is just a fancy way of saying maps that help oil and gas companies know where to drill). The customers pay a fee to access the data. It costs Pulse no money to add another customer. The business grows and grows, and the earnings compound and compound. They use the money to buy back shares, pay dividends, and add to the library.
But then you open up the company’s filings and you shake your head in dismay. The company had losses in 2009 and 2010 and eked out a small profit in 2007 and 2008. The profit and loss statement has something on it called participation survey revenue which actually isn’t revenue at all but expense reimbursement. And there is a huge expense called “amortization of seismic data library” that isn’t a true expense but a non-cash accounting fiction.
Luckily for our sake, the company uses a non-IFRS measure they call “Shareholder Free Cash Flow.” SFCF erases the non-cash charges and deducts the participation survey revenue from the earnings calculation. SFCF margins are enormous. The average SFCF margin over the past 10 years was 65%. The difference between the IFRS-sanctioned net income and the company-sponsored SFCF can be striking. Thus, Pulse trades at infinite P/E ratio as its last 12 months resulted in an IFRS loss of $2.3 million. But, it trades at only 10.2 times trailing SFCF (or a yield of 9.78% if you prefer).
It gets better. Customers pay the bulk of the cost for creating new maps, but Pulse is free to license the map to whomever they like (more on this below). With customers paying the bulk of the capex and an extremely high margin licensing business, Pulse truly is a company that drowns in cash.
Pulse Stock Performance vs. S&P 500, 2003-Present
Pulse’s library consists of 3D and 2D seismic data focused on Western Canada. Oil and gas companies need this data to identify where pockets of natural gas are. This data is particularly important for horizontal drilling, unconventional plays like shale and tight gas formations.
Pulse bears more than a passing resemblance to Sanborn Map Company, a prominent investment by a young Warren Buffett. Buffett invested a whopping 35% of his fund’s net assets in Sanborn. Sanborn was engaged in the preparation and sale of incredibly detailed maps of US cities. Their primary customers were 30 fire insurance companies.
The company’s moat held for 75 years. Per Buffett in his 1960 letter to his partners, during this incredibly long run, the company’s consistent profits was “accompanied by almost complete immunity to recession and lack of need for any sales effort.” Because the fire insurance companies were so dependent upon these maps and so irritated by Sanborn’s monopoly position, they took control of the company by placing several insurance executives on the board of directors. Even with their biggest customers controlling the board of directors, the company was still very profitable. Because Sanborn had very low capital needs, the company generated lots of free cash flow. Rather than returning it to the shareholders as dividends or share buybacks or reinvesting in growth opportunities, the feckless board of directors chose to accumulate an investment portfolio made up of roughly 50% bonds and 50% stocks. In fact, they even cut the dividend 5 times in 8 years.
Given the mean reverting nature of business competition, the inevitable occurred and the moat was breached when a new technology disrupted Sanborn’s business. Over a 20 year period, profits fell from $500,000 per year to $100,000 per year. Per Buffett again, “this amounted to an almost complete elimination of what had been sizable, stable earning power.” As Buffett was still in his “cigar butt” phase (this was a full 12 years before his See’s Candy investment), it is at this low point that he enters this picture. In the 1960 letter, Buffett states that when he bought in, the stock was at $45 and the investment portfolio alone was worth $65. Buffet came in as an activist to wanting to his hands on the investment horde. His agitation paid off as his shares were eventually exchange for the portfolio securities at fair value.
There are a few important lessons in the Samborn map story for moat hunters. First, it’s very good to be in the map business when the maps are difficult to replicate and vitally needed by the customers. Second, even a great moat, one that last for 75 years, can be undermined by bad management. Third, be wary of technological innovation that can disrupt the marketplace.
I favor relatively slowly evolving industries with few or no competitors or alternatives. The seismic data market certainly fits this description. We don’t anticipate any major new developments in mapping technology and the industry is effectively a duopoly.
Pulse’s chief competitor is Olympic, a division of Seitel Corp. Seitel is owned by ValueAct Capital, a well-known value investing firm. I should also note that Berkshire Hathaway made a bid for the company went it emerged from bankruptcy in 2003 but it was rejected.
There is also little customer concentration. No single customer represents more than 10% of shares.
Differentiated Business Model
Pulse’s basic business model is to build a database of proprietary information and license this data to businesses that have a vital need for the data. Pulse’s business model is based on repeatedly licensing and re-licensing their seismic data library. The library retains its usefulness and hence its cash generating value for decades. Further, it can be re-licensed at very little marginal expense for the company. Once the data is sufficiently monetized to cover the very low fixed costs, all additional income falls directly to the bottom-line.
Here, the subject matter happens to be seismic data but there are many other database companies such a Dun & Bradstreet, Gartner, Equifax, and Verisk. The basic economics are the same for all of these companies and they are popular with value investors. Once database companies reach a point of maturity, they can offer incredibly attractive economics because the business is highly scalable. When looking for high quality companies, this seems to be a fertile hunting ground.
A unique and highly attractive quirk of this industry is that customer pays 50-75% of the total cost of each participation surveys. Further, once the survey is completed, Pulse owns 100% of the acquired data and licenses it to additional parties on a non-exclusive basis. Such resales are unlimited in both time and amount and require minimal incremental cash costs, leading to a rapid payback period on new investments of typically less than three years and high returns thereafter. Pulse does not disclose maintenance capex but Seitel estimates it at just $25 million per year.
Someone on VIC likened it to a movie studio getting someone to underwrite the costs of a movie and then getting none of the box office or ancillary revenues. It’s hard to believe that it happens.
Does Pulse Have a Moat?
Quantitative Evidence of a Moat
When determining whether Pulse has a moat, our first step is to review the quantitative evidence. Our key metric for determining whether there is a moat is Return on Invested Capital (ROIC). Pulse’s ROIC is consistently very high. This makes intuitive sense as It takes almost no hard assets to run this business. Another key metric, steady and predictable free cash flow, is also positive as it has increased at just 7% CAGR over the past 10 years. One of the hallmarks of a great business is very little competition. As a result, I look at market share stability, a third metric. As discussed above, Pulse and Olympic/Seitl have control the Canadian market.
Qualitative Evidence of a Moat
The source of Pulse’s moat is a type of moat that we have not looked at previously: ownership of a unique asset. The nature of this moat is quite simple. The company holds proprietary data in the form of its seismic data library. This library contains mission-critical information for the oil and gas industry. The key, however, is that It would be very difficult for a competitor to replicate its library in a way that makes economic sense. There is little incentive for competitors to survey areas where Pulse already has data because the licensing cost is much lower than the costs of a new survey. In other words, it is dramatically cheaper for a customer to just license from Pulse rather than re-surveying the same area. Pulse estimates the replacement cost of their seismic data library at over $2 billion. Further, the moat is continuously being widened through participation surveys and acquisitions.
Pulse complements their proprietary data with deep domain expertise. They are a niche business and they don’t stray from their core competency. The president stated in his 2012 letter to shareholders, “We understand the seismic business, we love the seismic business, and we intend to stay in the seismic business.”
Two things I look for when determining whether a company has pricing power is (1) does their product or service serve a critical need of their customers and (2) is the price of their product or service a small part of the customers overall costs?
The answer for Pulse is “yes” and “yes.”
Pulse provides mission critical data to the oil and gas industry. As Mackie wrote: “With well costs running an estimated $8-15 MM/well for unconventional resource plays, the cost of developing prospective fields can run into the hundreds of millions. Given an oil & gas company could bankrupt itself drilling without the geological definition that seismic provides, they have virtually no choice but to use seismic data.”
The price of their product is only 5% of the total cost of a drilling project for the customer. This asymmetry gives Pulse the ability to raise prices without much worry. After all, the price increase will be an immaterial issue for the customer when reviewing the costs of the project but quite meaningful for Pulse.
High But Unpredictable Free Cash Flow
As discussed, the FCF margin is very high. It has averaged 65% over the past 10 years. What accounts for this high FCF margin and is it sustainable?
The company is adroit at dealing with the volatility. Pulse has very low fixed costs and a highly flexible operating model. They do not own any seismic survey equipment or directly employ any field personnel. This area is entirely outsourced. What few employees they do have are largely sales people who are paid on commission. This form of compensation helps to match costs to actual sales. These features results in lower cash flow volatility than would otherwise occur by enabling them to respond quickly to changes in demand.
Further, because the business didn’t really require incremental invested capital, it found itself generating gobs of cash. As stated above, the customers foot the bill for most of the growth capex and maintenance capex is assumed to be very low.
So the margins seem sustainable, but the results are very lumpy and tied to a large extent to the commodity cycle. SFCF has increased at a CAGR of 7% over the last 5 years and 17% over the last 10 years. The growth is certainly reasonable but not crazy. I typically prefer companies with a high level of recurring revenue. Recurring revenue makes it easier to project future cash flows and reduces risk. Pulse’s cash flows are anything but predictable. The company acknowledges this issue. “The first quarter’s remarkable upside demonstrated the innately poor visibility of seismic revenues. Data library sales can occur in large lumps and are driven by factors beyond standard industry metrics like commodity prices and drilling rig utilization,” wrote the CEO in the 2012 annual report.
As seen in the quote above, the company argues that Pulse’s results are largely decoupled from the commodity cycle. At first glance, this sounds strange. How could a provider of services to the oil and gas industry not be impacted by the ups and downs of commodity pricing? I am somewhat skeptical, but there are several growth drivers that are independent of the commodity cycle.
First, because Pulse is in such a strong bargaining position with its customers, not only can it require them to fund 50-75% of the cost of new surveys, the licenses Pulse grants are not freely transferable. This feature helps to drive additional sales. For example, the initial land lessee might look to bring in a partner to diversify his risk, inject additional capital or to generate some cash. The license to the Pulse data that the initial lessee has is not transferable to the new partner. Pulse refers to such deals as “transaction-based sales” and says they comprised a significant portion of 2012 revenue.
Pulse also gets a helping hand from the Alberta government. The vast majority of Pulse’s seismic data relates to land in the province of Alberta. Alberta awards the rights to engage in oil and natural gas exploration and development on specific land to specific companies. These rights expire typically after 5 years but sometimes after 3 unless the company begins drilling. This tight deadline drives demand for seismic data. Similarly, rather than letting rights simply expire, companies engage in transactions to retain partial interests in their land but let other parties begin drilling. Such transactions drive further demand for seismic data since the license Pulse granted the original party is not transferable to the new party. Any partner or acquirer wishing to use the data will be required to pay a fee.
Beyond that, Pulse has sufficient cash flow to invest in high return capital projects. Due to the excellent results in 2012, they have more cash than ever before. Further, Western Canada is the site of promising unconventional oil and liquids-rich gas plays.
Management and Capital Allocation
High cash generation is meaningless if management does not put it to good use. Luckily management at Pulse is much better than at Sanborn Map. The Chairman of the Board is Robert Robotti. This name should be familiar to value investors. Robotti is the principal at Robotti & Company and a well-known value investor. With Robotti at the helm, capital allocation has been excellent.
Not surprisingly, the capital allocation has been perfect. The company buy back 10% of shares each year, which is the maximum allowed under Canadian law. The company also pays a dividend of $.08 per share.
Pulse’s largest acquisition to-date has been a huge success. In 2010, Pulse bought a data library from Divestco for cash of $55.6 million and 14.3 million newly issued shares. As of September 2013, Pulse had generated a total of $63.4 million in sales (114% of cash purchase price) and bought back 59% of the shares.
Most moat discussions focus on the long-running moats of big cap companies like Coke, Wal-Mart and P&G. But, identifying smaller companies with durable competitive advantages and skillful management teams can be more profitable. After all, smaller companies often after longer growth runways. Because they are small, they have the ability to compound capital at high rates of return over long periods of time.
That’s why I appreciate Mackie bringing Pulse to our attention. While I love the business model and the cash-generation, I am not as enthusiastic about this company as Mackie because of the lumpy results. The uneven results make it too difficult for me to project a forward growth rate. Nevertheless, Pulse provided a fascinating look at a very interesting business model – one that we can draw from in the future.
 Because Pulse is Canadian, all references are to Canadian dollars per IFRS unless otherwise noted.
 Here, ROIC = FCF / (Net Working Capital + Net Fixed Assets). I am using FCF rather than EBIT because, as described below, the accounting for this company and this industry is rather screwy. It creates too much of a distortion to use EBIT per GAAP without any adjustments.
 Q3 2013 Corporate Presentation