[ELLI – Ellie Mae] SaaS With Class…But Still Too Expensive

Several weeks ago, the stock of Ellie Mae, a SaaS that automates the grueling task of underwriting residential mortgages, got hammered when the company reported revenue weakness on lower than expected industry refinancing volumes and longer than expected ramp times for larger enterprise customers, prompting management to dramatically cut its full year revenue growth guidance from ~21% to just ~12%.  While these factors hit the 40% of ELLI’s revenue tied to transactions, the remaining 60% of its business, contracted revenue invariant to origination volumes, expanded by 35% y/y, continuing its robust growth trend.

Mortgage volumes will do what they do in any given year.  But what won’t change is the structural complexity of underwriting a mortgage.  From the potential home buyer first reaching out to a loan officer; to the loan officer pulling a credit report and discussing pries and terms with the borrower, and ordering an appraisal, title insurance, and other services from third parties; to the underwriting department assessing compliance with regs and underwriting guidelines; to the closing department preparing the bundle of signature papers; and multiple departments reviewing all the docs for fraud and compliance and accuracy along the way before the loan is finally sold to an investor, over a thousand pages – application, credit, flood, title, borrower financials, property appraisal, fraud, compliance, insurance – are spawned by a dozen or more service providers operating across silo’ed databases and systems.  Jerky spurts of information are manually pushed and pulled between multiple parties embarking on multiple workflows.

The mortgage manufacturing process is a massive source of frustration for the home buyer, who is hounded by the loan officer for supporting documents or reminded about missing information that he is sure the bank already has, often weeks after he’s submitted the relevant forms and mentally prepared for the next stage.  And at the last stage of the production process, when the funded mortgage is ready for sale, the lender must now wait, typically up to 30 days, for the mortgage investor to image, extract, and index all the information from the PDF’ed forms that have mounted up to this point before finally receiving the sales proceeds with which to then underwrite the next mortgage.  This broken, inefficient process has been crying for the automation that Ellie Mae’s SaaS delivers.

A mortgage underwriter’s chief concern is to underwrite as many loans as responsibly as possible, for as cheap as possible.  But every day that paper gets shuffled to and fro because of overlooked data fields and unticked boxes and other dropped balls, is another day that the mortgage gets burdened with operating costs.  The industry is also contending with the stricter underwriting and processing standards – including more onerous documentation requirements and greater use of unbiased third party service providers – demanded by regulators and investors, contributing to a steady rise in the cost of producing a mortgage from ~$4k/loan in 2009 to ~$9k today (per the Mortgage Bankers Association).

Unlike many vertical SaaS companies, who are shifting from providing proprietary software to increasingly fostering an ecosystem of industry constituents, Ellie Mae began as a network and backed into software.  After 3 years of development, in late 2000 Ellie Mae launched what is today called the “Ellie Mae Network”, which connects residential mortgage originators and lenders to service providers and clips fees from transactions among network participants.  In the following years, the company acquired 2 loan processing and document prep software companies and launched its flagship Encompass in 2003, augmenting its solution bundle through various acquisitions over the years.

Today, the Encompass SaaS is tightly integrated into the workflows of over 200k subscribers (96% of whom renew every year) who use it to manage the paperwork and to-do items for nearly all the links in the origination chain – loan processing, underwriting, funding, application prep, disclosure agreements, closing docs, regulatory and investor compliance, record management, customer communications – through a single web portal.  So, for instance, rather than email the borrower PDF’ed documents to print, sign, scan, and email back, the loan officer can simply direct him to a web link, whereupon logging in, the borrower can complete a loan application, e-sign documents, and upload financial statements and other required documents that will then all be automatically and sanely organized for the loan officer and other bank personnel to review.  Encompass accounts for 60% of Ellie Mae’s total revenue (mostly recurring monthly subscription fees, but also Success Based Pricing tied to funded mortgages), with the balance coming from transaction fees paid by service providers and investors plugged into the Ellie Mae Network, and to a much lesser degree consulting, education, implementation, and analytics services.  Encompass and Ellie Mae Network have always been intertwined, with Encompass acting as the portal through which mortgage originators access service providers, for whom Ellie Mae is typically the most productive distribution channel.

But as many SaaS/platform-fused companies have figured out, next-level scaling means relinquishing absolute control over application development and instead offering tools for customers to build what they need.  So the company is providing 2 distinct sets of APIs on its Encompass Lending Platform that it is rolling out this and next year.  The first set allows customers to build their own applications or augment Ellie Mae’s existing software suite.  The second lets service providers (offering title insurance, mortgage insurance, property appraisals, credit checks, flood certifications, etc.) integrate their applications into the Ellie Mae platform [Ellie Mae vets these applications for quality and security the same way Apple quality controls for apps in its App Store], replacing batched data dumps with real time data feeds to lenders and investors who are plugged into the network.

Given Ellie Mae’s mature 35% share of retail origination volume [the company says it’s 30%, but 15% of the TAM is represented by large money center banks that, for now, are largely outside ELLI’s reach], the dis-aggregation of applications into modularized APIs that allow custom built software to call on generic workflow engines residing at the platform layer, is a necessary first step to expanding beyond residential mortgages, to lateral loan categories underwritten by the company’s bank and credit union customers – personal, commercial, student, small business, and auto loans.

But Ellie Mae’s share of residential origination volume understates its promising monetization potential.  As previously noted, the production cost of a typical mortgage is close to $9k.  By reducing the time to carry a loan to close by 20 days, an active, full bundle Encompass user can shave $2k-$3k off, a huge savings compared to the ~$400/loan paid to the company for core origination ($125), network ($135), and currently available add-on services ($140) [but does not include $225/loan from a slew of near-term potential add-ons nor unspecified revenue from future initiatives like business intelligence and other machine-learning fueled analytics that are currently in the works].  By contrast, the current average Ellie Mae user pays just ~$145/loan.  The company has credibly demonstrated its ability to extract incremental unit revenue atop subscriber additions, growing, for instance, revenue per 2012 vintage customer by 16%/year from 2010-2016 and contract revenue per active user by ~30%/year over that time.  If we apply the upper bound of per loan revenue against the addressable volume of loans, I think the dollar TAM opportunity is more like $4bn, so ELLI has ~10% share.

Still, even if we grant that the addressable market is larger than volume share would imply and that Ellie Mae, by virtue of its network effect + fat solutions bundle, is positioned to profitably capture share, the stock’s valuation still renders it an unappealing investment.  If we assume that the housing market mostly cooperates, and that ELLI’s revenue compounds by 26%/year for the next 7 years (vs. 20% ytd and management’s long-term expectation of 25%) while layering in the operating leverage that gets EBITDA to grow by 29%/year, we’re at $2bn in revenue and $540mn in EBITDA, which trickles down to ~$5.50 in cash earnings per share (EBITDA less capex less taxes less capitalized software development costs divided by shares).  In this scenario, ELLI has claimed just under half the residential mortgage $TAM and is a mature, dominant concern, whose fate is increasingly tethered to the housing cycle.  Even if we apply a generous 25x year 7 earnings multiple and tack on the interim accumulated cash + today’s cash, we can do no better than $168/share, a 10% IRR off the current price.