Ellie Mae is one of the few SMID-cap SaaS vendors I track whose stock has lost value over the last year, its growth narrative clouded by waning growth in mortgage originations. During the first half of 2018, a common view espoused by housing-related companies was that refinancing volumes had pretty much nowhere left to go but up and purchase volumes – far less sensitive to rates and fueled by a secular tailwind of millennial demand(1) – would continue growing by mid/high-single digits for the foreseeable future despite tight inventory conditions. But things have been weaker than imagined, with First American and Fidelity National experiencing decelerating purchase title orders and refinancing volumes continuing to plummet. Ellie Mae’s management was caught off guard and after tempering full year revenue guidance in mid-2017, was forced to do so again, several times, in 2018.
Rather than charge its customers a fixed monthly fee, Ellie Mae engages in “success-based” pricing, where lenders pay a base fee up to a certain number of loans plus an additional amount for every closed loan above that threshold (regardless of the actual funded amount or whether the loan is part of a purchase or refi transaction). The revenue model puts Ellie Mae in partnership with its customers (since they both prosper from growing loan volumes), but exposes it to cyclicality that most SaaS vendors avoid. Weakness in closed loan fees and revenue from other volume sensitive services, which account for 35% of Ellie Mae’s total revenue, dampened Ellie Mae’s overall growth rate in 2017. [closed loan fees, over 20% of total revenue in 2016, plummeted by 27% even as contracted base fees have compounded by >35%/yr over the last few years. Contracted revenue (~70% of total), which includes aforementioned base fees plus services and subscription revenue on various ancillary products, has expanded by 65% since 2016 while non-contracted revenue, success-based fees and fees from transactions on the Ellie Mae Network, has contracted by 12%.]
Anyway, let’s set the macro backdrop aside and get into why Ellie Mae even matters. Underwriting a mortgage is a complicated process buffeted by regulations and inefficiencies, as described in my last ELLI post:
“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”.
To address escalating regulatory demands, lenders have, for the most part, simply ratcheted up their hiring. Two-thirds of the $8k spent in taking a loan from application to close is concentrated in personnel…that is, paying lenders and back office employees to verify and reconcile information; procure services; and wait around for someone upstream in the production chain to complete his part of the paper work.
The mortgage production process has several phases, with a dozen to-dos falling under each phase. The green boxes in the below exhibit represent tasks that can be automated.
Source: Ellie Mae
Ellie Mae’s flagship product, Encompass, mostly addresses the 44 days that pass from when the mortgage application is received to when it is funded. As the loan makes its way through this pipeline, it is scanned for defects at various points so that it emerges from the funnel in such pristine condition that the investor who buys it won’t have to kick it back to the originating lender because of some overlooked defect(2). Between the core SaaS product that automates these processes and stores all related records in a single system, and the Encompass Network through which Encompass users transact with third parties who verify income and employment, conduct property appraisals, and underwrite mortgage, flood, and title insurance, Ellie Mae thinks it can cut a process that typically takes over 70 days to something closer to 15. In doing so, the lender conserves value otherwise absorbed by personnel costs, and reduces both the interest costs of carrying a mortgage on a warehouse line and the opportunity cost of mortgages that can’t be underwritten in the meantime. Once a system like this gets set-up, with lenders and back office employees relying on Encompass as the operating system of their mortgage manufacturing process, it is very difficult to dislodge. Ellie’s customer retention rates are 95%+.
At the center of everything is the Encompass Lending Platform, a set of tools and APIs that Ellie Mae uses to build proprietary software, including the core Encompass SaaS product that makes up 60% of its revenue(3). These APIs are also exposed to customers (who use them to build Encompass customizations) and to third party service providers (who use them to access lenders on Encompass’ platform)(4). While network effects are ostensibly apparent in the bidirectional pull between service providers and Encompass customers, the whole digital mortgage space is replete with non-rivalrous integrations – service providers basically see Ellie Mae, Black Knight, Roostify, etc. as distribution channels; plugging into multiple platforms is nbd – and everyone seems to offer their own networks or exchanges. I mostly see the Ellie Mae Network as table stakes, rather than a differentiating advantage. That said, with ~40% of mortgages processed on its platform, Ellie Mae is certainly a required integration for service providers and mortgage investors (that’s why they, and not the lenders, pay the transaction fees). However, I don’t think the dependence is mutual. For a lender, does an network with 100 appraisers really offer much more value than one with just 20?
There are a few different ways to think about the addressable market and Ellie Mae’s share of it: the company supports ~2,400 lenders vs. ~6,000 lenders nationwide, including 3 of the top 10 and 10 of the top 30; hosts 200k+ active users against a a nationwide pool of ~850k lending officers and immediate support personnel; and processes 2.5mn loans against ~9mn mortgages originated  in the US over the last 12 months. But the more effectively Encompass does its job, the more automated the origination process becomes, and the fewer the number of loan officers required per loan. And as mentioned earlier, Ellie Mae’s revenue is in part tied to loan volumes.
Thus, management has been encouraging investors to frame the company’s market share in terms of mortgage volume and its monetization potential in terms of revenue per mortgage. Ellie Mae extracts around $180 per originated loan against a potential ceiling of close to $400 based on the services that are presently available on its platform (base fees + per loan closing fee = ~$100; Ellie Mae Network (flood reports, income verification, etc.): $130; other software: ~$140). Ushering a loan through the origination funnel might cost $8,000+, while the potential cost savings of adopting the Encompass platform – namely, lower personnel costs as the time to originate compresses from over 40 days to ~10 to 15 – are substantial at $2,600, 20%-30% of which might eventually be shared with Ellie (per management). The cost-benefit gap is probably wide enough to accommodate several points of annual price hikes.
Here’s the bull case. Multiplying $2,600 by ~9mn funded mortgage loans yields around $20bn-$25bn in industry-wide cost savings. Assuming 30% of that is shared with software companies like Ellie Mae who enable those savings gets us to a total addressable market of ~$7bn(5). Ellie Mae’s ~$460mn of revenue accounts for ~7% of that.
So, assuming that Ellie Mae can grow its revenue per loan by ~10%/year and walks its market share up to 50% of originations, we would be looking at $1.6bn in revenue, which drops down to ~$450mn in EBITDA (including stock comp expense) assuming 28% margins (vs. 17% margins today and Black Knight’s margins of 46%). Blow out the shares by ~5%/year and I can pencil out ~$4.30/share of cash earnings (which deducts capex and capitalized software). Slap a 25 multiple on that, tack on the accumulated cash, and there’s your ~12% return.
But there are a few competitive entanglements to consider. Recently, Ellie Mae has been expanding its presence at the front end of the value chain, with tools that automate marketing campaigns (Encompass CRM) and automatically assign leads to the right loan officer (Velocify; acquired in October 2017 for $128mn). With Consumer Connect, Ellie Mae white-labels a web-based application through which mortgage borrowers can research loan options, upload documents, obtain status updates, access loan officers, and upload documents that then flow automatically into Encompass. Ellie Mae offers Consumer Connect for free to its customers in the hopes that doing so will drive more mortgage volumes into Encompass.
This adjacent move is a reaction to the rising prominence of various upstarts in the digital mortgage application space, namely Roostify and Blend. Whereas Ellie Mae began by directly targeting the hidden belly of the mortgage origination process before now forward integrating functions that are closer to the consumer, Roostify and Blend, founded in 2014 and 2012 respectively(6), began by interfacing with borrowers before extending into the nuts of bolts of origination. Both have now secured major banks like Chase and Wells Fargo as customers, and are weaving up- and downstream integrations with other relevant players in lead generation and mortgage servicing.
These front-end competitors are becoming a critical point of integration, pulling ever more utility into earlier stages of the value chain, away from Ellie Mae’s ecosystem. Roostify integrates into LendingTree so that a LendingTree lead who selects Bank A for a mortgage loan will immediately hop into a Roostify application that is white labeled and branded by Bank A. Roostify (and Blend) also both integrate into Encompass, so if Bank A happens to rely on Encompass to streamline mortgage originations, all the relevant data from Roostify’s application is delivered into Encompass. But Roostify and Blend touch the borrower earlier than Ellie Mae and it strikes me that many of those green boxes in the earlier exhibit either currently are or could just as easily be handled by them, including income/asset/identity/employment verification. Much of the inefficiency in the traditional underwriting process is exorcised by synchronously matching the delivery and receipt of information and storing it all in a single place so there isn’t a back and forth shuffle of pdf’s. A digital mortgage application already supplies these functions, either natively or through integrations with third various partners like Yodlee, Equifax, docutech, Salesforce. And so, Roostify and Blend are predictably positioning themselves as “hubs” through which all underwriting constituents – consumers, lenders, service providers – can converge to coordinate tasks and get real time status updates on the loan’s procession through the pipeline.
Black Knight (BKI), a mortgage servicing platform spun off of title insurer Fidelity National last September(1), is another competitor to consider. [Mortgage servicing involves collecting and recording mortgage, tax, and insurance payments on a mortgage, and comes after the origination phase. Servicing can be done by the underwriting bank or by a third party like Ocwen]. Black Knight is a giant in the realm of software/data/analytics for the mortgage niche: it supports $8.5bn of enterprise value on $1bn of revenue and claims 20 to the top 25 mortgage servicers as customers. Around 70% of outstanding US mortgages are serviced on its platform. But in addition to mortgage servicing, Black Knight offers mortgage origination software through its Empower! brand(7). Empower! is only ~40% of Ellie Mae’s size and at less than 20% of Black Knight’s total revenue, plays second fiddle to the company’s servicing business. But management seems to have taken a renewed interest in reviving this franchise, as evidenced by recent product enhancements and new financial disclosure that separately discloses its revenue. Moreover, last year the company launched a mid-tier version of Empower! (Empower Now!), extending its reach to regional and mid-market lenders where Ellie Mae operates. Ellie Mae, meanwhile, is migrating upmarket, where Black Knight is more firmly entrenched, possibly setting the two companies on a collision course.
But in its attempts to leverage its dominant position in servicing to an expanded presence in originations, Black Knight faces several obstacles. First, origination and servicing are distinct processes with little overlap, steered by distinct sets of employees. There doesn’t seem to be any functional logic to buying a bundled product that includes servicing and origination software. In terms of bundling, it seems Roostify and Blend have a more compelling vector of attack: the loan officer responsible for managing leads and reviewing digital applications also takes point in walking that mortgage through to funding. But his responsibilities certainly don’t extend to collecting payments. Second, Black Knight is not really offering a modern SaaS product – with real time updates, rapid product improvement rollouts, and a common set of capabilities available to all customers – so much as it is standing up an expensive, customized (and some might say, outdated) system that can take up to a year to implement. Unlike Ellie Mae, you will not hear Black Knight talk about migrating infrastructure to AWS or building off a modular service oriented architecture or harping on APIs that customers and ISVs can use to extend the core product.
A bespoke solution loaded with professional services may be appropriate for large banks with $100bn+ balance sheets – and indeed, Ellie Mae is losing to competition among the few hundred lending organizations that make up the high end of the market – but a lighter weight version of this arrangement may have a harder time gaining traction for smaller banks than a streamlined SaaS product. Also, the cost structure required to support a high end solution are not amenable to a lower priced instantiation, and even Black Knight is limiting its purview for Empower Now! to the top 200 lenders (out of a universe of 5,000+). It is generally easier to get organizational buy-in to gradually nudge a “good enough” product upmarket than push a natively enterprise solution downmarket.
Show me the money
Here are the components of Ellie Mae’s revenue growth:
Average active users and rev/active user have decelerated meaningfully over the last few years. Even so, 20% ytd revenue growth in contracting origination market ain’t bad!
At least one sell-side analyst has praised the company for protecting margins in these trying times. But, whether the company deserves praise in this regard depends on your definition of profits.
So do you look at:
(a) EBITDA ex. stock comp? ($mn)
(b) EBITDA less stock comp?
(c) EBITDA less stock comp less “acquisition of internal use software” (capitalized software)?
(d) EBITDA less stock comp less “acquisition of internal use software” less capex?
You know my thoughts on stock based comp. I think it’s okay to back it out of profits so long as you blow out the shares and evaluate things on a per share basis. In other words, the cost of issuing stock must be reckoned with somewhere. The problem is that Ellie Mae’s adjusted EBITDA and adjusted EBITDA margins account for it nowhere. As measures of profitability, they are wrong, full stop.
But (c) seems like a reasonable measure of “real” pre-tax cash profits, since capitalized software is functionally equivalent to either R&D or infrastructure support costs. And the trend here is ugly. Ellie Mae’s revenue has increased by $375mn since 2012; profits have declined by $14mn during that time. There’s been no leverage on any part of the cost structure over the last 6 years (except maybe a little bit on G&A):
But on the other hand, I don’t know why we should expect otherwise. In pursuit of a lucrative market, Ellie Mae has, since 2013, been investing aggressively in sales, customer support, product development, and technology infrastructure(8), and isn’t that what it should be doing? On the other, other hand, maybe the competitive environment has intensified to such an extent that the recurring table stakes investments required to stay relevant is trending higher and higher. I suspect both explanations are true.
(1) The largest part of the millennial population, those between the ages of 26 and 27, are just a few years away from the age at which they start buying homes.
(2) Ellie continuously keeps up with all the regulatory minutiae at the national and local level.
(3) The other 40% comes from a/ fees that Ellie Mae earns on transactions that take place on the Ellie Mae Network, b/ professional services, and c/ various other software and data solutions (closing, CRM, loan compliance, access to an online database of Fannie/Freddie lending guidelines and forms, etc.)
(4) As I’ve written before, monolithic software applications are being disaggregated into a set of specialized functions that communicate through APIs, enabling speedier, more scalable, more flexible product development
(5) You could make the case that this figure, which only takes into account the saved personnel costs, understate the actual TAM. If you include the $670 that is spent on technology for every loan origination, the real opportunity might be closer to $13bn.
(6) Blend has raised $160mn in venture funding over its short life and is now valued at ~$500mn.
(7) Empower! has been around since at least 2008, back when Black Knight was known as Lender Processing Services.
(8) While the company is leveraging AWS for newer products, including private data lakes, Ellie Mae, for the most part, operates and maintains data centers that host software delivered on-demand to its customers.
Disclaimer: As of the time this report was posted, accounts managed by Compound Insight held shares of ELLI. This may have changed at any time since.