When David Cambria sat down with Eric Elfman to discuss his willingness to try Onit software, he stated that if ADM in-house lawyers were required to engage “in a single unnatural act,” the implementation would fail.

Cambria elaborates, “Why are we all so comfortable with Word, Excel, and Outlook? Because these tools don’t have an opinion about how we do our work. Enterprise software, however, always has an opinion.”

Hardened by 25 years of work experience in consulting and legal operations, David communicated his need for workflow tools that did not require his lawyers to change. Further, he needed significant productivity gains and a steady stream of clean, reliable data to better manage the department. A high bar for success.  Yet, according to David, Onit managed to deliver.

Cambria, Global Director of Legal Operations at ADM, recounts this story during Week 6 of “How Innovation Diffuses in the Legal Industry.”  Eric Elfman, Founder & CEO of Onit, was also present, giving his own entertaining version of a project that went on to win a 2017 ACC Value Challenge Award.

By inviting Cambria and Elfman to class, I hoped students would get a glimpse into the type of buyer-supplier relationship that enables a legaltech company to successfully “cross the chasm.” See Posts 024026 (discussing chasm framework, its connection to diffusion theory, and its applicability to the legal industry).


For a summary of Week 2 guest lectures (Pangea3, Practical Law Company, Hotshot), see Post 032. For week 3 (consultative sales at Thomson Reuters), see Post 034. For Week 4 (a deep dive into Axiom), see Post 036. For Week 5 (law firm examples of intrapreneurship), see Post 039.


Crossing the Chasm

I knew I hit pay dirt when Elfman came to class with a dog-eared copy of Crossing the Chasm.  Naturally, I had to ask, “Have you ever crossed the chasm?”  With an enormous grin, Eric replies, “Twice.”

The first time was with Datacert, an e-billing company Eric founded in 1998 with $1,000 of his own money.  The timing and concept were right, as Elfman quickly landed five Fortune 500 clients, making it relatively easy to attract investor money to build out the product and scale. When Eric left the Datacert in 2008, it was valued at $60 million. In 2014, Wolters Kluwer acquired Datacert for $290 million, merging it with TyMetrix to create what is now known as Wolters Kluwer ELM Solutions.  (The acronym “ELM” stands for enterprise legal management.)

The second crossing was with Onit, a business process automation company Elfman founded in 2010. This time, Eric put $1 million of his own money followed by four rounds of outside investment (a mix of debt and equity) totaling $16.4 million.  Eric stated that the company crossed the chasm approximately a year ago when operating income could more than cover ongoing R&D and sales efforts.  “That is not to say we won’t raise more money,” added Elfman. “Simplicity is extremely expensive to create. You also need to have high quality products when customers want to buy them.”

Onit’s core product is configurable software that can be deployed relatively cheaply and pointed at a wide range of legal department needs.  Established applications include legal spend management, matter management, contract management, legal holds, legal service requests, NDAs, and virtually any type of work flow involving knowledge workers.

Onits’ major competitors are enterprise software providers that serve corporate legal departments. However, most competitor offerings are built around a single problem. This means that legal departments tend to have several enterprise systems that can’t talk to each other very well. As discussed in more detail below, legal departments are perennially underwhelmed with their enterprise software incumbents (my observation, not Elfman’s).

Onit currently has 105 employees in the US, UK and India, and $10 million in annual revenue. According to Elfman, for the last three years, the company has been growing at a 50% annual rate.

Corporate legal departments as a target niche market

As I listen to Cambria and Elfman share their experiences, I am surprised by how well the narrative fits the crossing-the-chasm framework.

To refresh readers’ understanding, a company starts life with a generic product that likely impresses technology enthusiasts but lacks the features needed for broad mainstream adoption. Thus, to cross the chasm and achieve commercial success, a company must (a) target a niche market that could benefit from the innovation, (b) identify its biggest pain points, and (c) work backwards to build a “whole product solution” that becomes the “the only reasonable buying proposition” for the target market customer.  Moore, Crossing the Chasm (1st ed. 1991) at p. 110; see also Post 024 (summarizing basic framework).

This is Moore’s “big fish, small pond” strategy, which is designed to create focus on the narrow set of clients and conserve the bandwidth of key personnel.  See Post 025. If executed properly, the post-chasm company has successful commercial relationships with “pragmatist” mainstream customers. This sets off a word-of-mouth campaign that dramatically reduces the cost of sales. Further, once inside the mainstream market, the company is well-positioned to develop and sell future products and services.

In short, crossing the chasm is a one-time event that changes everything for the better. See graphic below:

Well, what is Onit’s target niche market (or small pond)?  Here I get an important lesson in framing.

Virtually all legaltech companies target a discrete problem or complex task that exists within a legal department. These problems or tasks include e-billing, matter management, document management, e-discovery, contract analytics, etc.  When evaluating this market structure, the natural capitalist impulse is to integrate these disparate systems into a single enterprise solution, thus achieving economies of scope and scale. Indeed, this is the logic behind many legaltech acquisitions, including the Datacert-Tymetrix tie-up. Framed in this way (which is the way most legal insiders see legaltech), the small pond is one or two significant problems or tasks inside a legal department.

But that is not Onit’s strategy.  Onit is a business process automation company where legal departments are viewed as a small but influential beachhead that can provide access to rest of the corporation. Thus, the addressable market is not all corporate legal departments (which might be $3-5 billion), but corporate knowledge workers struggling to collaborate effectively within and across business units (probably 100x bigger). Framed in this manner, the small pond is legal department operations.

Few tech entrepreneurs would be anxious to have legal as their initial target market. The field is highly technical; the clientele are demanding; and the financial upside is limited. But Elfman sees things differently.  “The lawyers are the laggards. They are the Department of No. If we can win them over, the rest of the corporation is a lot easier.”

I am inclined to take Elfman seriously because he and his team are obsessively focused on delivering a whole product solution. To fully grasp what this means, we need to understand Onit as compared to its primary competition.

Compared to what?

In Crossing the Chasm, Geoffrey Moore makes the point that prospective clients are unwilling to strain their attention span to hear your pitch. Thus, a product needs to be positioned against what is familiar and established, thus enabling target clients to quickly categorize your product.  Yet, to generate curiosity and interest, the product also needs to be different in a way that delivers a substantial benefit. See pp. 159-61.

As previously noted, Onit’s primary competitors are enterprise software companies that offer solutions to one or more legal department needs, such as e-billing, matter management, contract automation, or data analytics. In my travels to various industry events involving legal technology, I often hear the refrain, “Everybody hates their e-billing vendor.”  The same tends to be true for document and matter management. To date, no company has emerged as the obvious first choice.

Most of these companies got their foothold many years ago when legal departments were growing rapidly and general counsel and their lieutenants felt vulnerable regarding the lack of basic systems and controls. For example, without enormous manual effort, the department could not answer basic questions related to outside counsel spending; or the department couldn’t generate a useful status report on pending litigation; or lawyers struggled to locate prior work product. In each case, there was an enterprise software solution or platform designed to make that problem go away.

Indeed, Elfman tells the story of how he got the idea for Datacert. After completing his MBA at Rice in 1995, he went to work for a litigation consulting firm that specialized in forensic accounting.  While working on an engagement for Exxon, Eric asked the head of litigation about the size of his total annual spend. The AGC responded, “I’m not sure.  Somewhere between $200 and $400 million.”

Elfman describes this exchange as “the moment that changed my life.”  The business opportunity was large and obvious: use technology to apply basic accounting discipline to corporate legal spending.

Datacert and Elfman were extremely successful making sales to a lot of large corporations. Eventually, Datacert would land 130 companies in the Fortune 500, including #1, #2, #3, and #5.  Yet, Datacert also became part of the large cadre of enterprise software companies that legal departments complain about (this observation is based on my own industry knowledge, not any comments made by Elfman regarding his former company).

Root cause

As I listen to David Cambria and Eric Elfman discuss their collaboration, a deeper understanding of the problem comes into focus.

As David points out, when enterprise software is pointed at a specific problem, it develops a strong opinion about how the work should be done. Invariably, that opinion adds steps to the workflow, often without delivering any immediate or tangible returns to the worker trying to do their job. Naturally, people being people, they find ways of minimizing their interaction with the system. Thus, the resulting incomplete and uneven usage undermines the value of the enterprise solution. It also limits — possibly to zero — the amount of usable data the system produces.

In theory, management can fix this problem by mandating usage.  They can fire people. They can reduce or withhold bonuses.  Political capital, however, is limited.  Few bosses want the troops grumbling about how a six-figure software mistake is hindering their ability to do their jobs. So the natural equilibrium becomes enterprise software that is half used. This is usually a modest improvement over the prior state of affairs, but well short of expectations when the licensing agreement was signed.

This recurring cycle explains why David Cambria has such disdain for business solutions that require unnatural acts. Likewise, this is why Eric Elfman was ready to leave Datacert after ten years at CEO.  This was a game he could not win.

What problem is Onit trying to solve?

Eric Elfman left Datacert in 2008.  Two years later, he started Onit with Eric Smith, Datacert’s longtime CTO.  Yet it wasn’t until 2011 that Elfman and Smith came up with the core idea for Onit, which is “collaborative process automation for knowledge workers.”

Not very intuitive, right?

To Geoffrey Moore’s point, it is very difficult to understand an innovation without one or two familiar reference points. This is particularly true with something as abstract as software. Thus, the graphic below proved to be enormously useful to the class.

On the left side (in green) is enterprise software, which attempts to solve problems through top-down controls.  Although these solutions tend to be complex (requiring IT support) and expensive (big up-front fees and implementation), they hold out the promise of permanently eradicating a serious problem. The implicit assumption is that workers will use the system as designed — an assumption that, experience shows, is often unjustified and unrealistic.

On the right side (in orange) are Enterprise 2.0 tools (like Slack, Zoom, or Yammer). Individual users and work teams like these tools because they increase the velocity of employee communication.  Corporations are happy to support Enterprise 2.0 tools because they are cheap and low risk. But they also don’t produce any structured data that senior managers need to assess and improve organizational performance.

Despite billions of dollars spent on enterprise software and the hype and popularity of Enterprise 2.0, Elfman observes that “virtually all knowledge work and processes are executed outside of these systems.” Instead, in most organizations, workers try to do everything with familiar Microsoft tools:

  • Email is the intake and “collaboration” platform, within and across business units
  • Word documents are the “forms” solution
  • Excel is used for tracking and reporting
  • Sharepoint is used as a document repository

Virtually all legal operations professionals will acknowledge that these tools are breaking down as solutions. They are just not fit for purpose.

Onit (in blue) is trying to fill in the middle ground between Enterprise (green) and Enterprise 2.0 (orange). The key innovation of Onit is that it enables a business process owner to work backwards from how people work (people-centric) rather than backwards from an acute organizational pain point (problem-centric) and thereafter expecting workers to get onboard.

“Bring the work to the people”

When Cambria signed on with Onit, he had a vision to “bring the work to the people.” Where are the people in ADM’s legal department? Probably somewhere near a device where they read their email.

Onit is behind a wide range of automated workflows at ADM, including: (1) matter intake and routing, (2) early case assessments, (3) liability reserves, (4) invoice review and approvals, (5) settlement authority requests, (6) recording of matter disposition, and (7) on-demand NDAs. Yet, for most ADM lawyers, Onit is barely visible:  it’s all point-and-click tasks and hyperlinks embedded inside emails — highly natural acts for lawyers. Cf. Post 040 (per “lawyer theory of value,” lawyers have a strong preference to be left alone to do legal work).

Cambria or a member of his staff are usually the “business process owner” for each of these processes.  Onit is simple and flexible enough for them to do a fair amount of programming on their own — no need to involve corporate or department IT. This is ideal because the legal ops team is close enough to the work to gauge what the workforce is willing to accept. And If they are wrong, adjustments can be made cheaply and quickly.

Nudges and the Onit backend

One way that Cambria drives the broader agenda of the department is to include “nudges” in the Onit workflow.  A nudge makes it modestly more difficult for lawyers to override an established playbook solution. For example, if an ADM in-house lawyer wants to retain a law firm that is not on ADM’s preferred panel list (ADM winnowed 700 law firms down to a preferred provider list of 20, see “How ADM Cut Its Outside Counsel Rosters By 680 Law Firms,” Law360, June 8, 2016), a text box appears that requires a written explanation.  Because this choice requires additional work and invites scrutiny from the boss, it is chosen less often.  Explains Cambria, “I’m always mixing the peas in with the mashed potatoes.”

Although Onit is largely invisible to a substantial portion of the ADM legal department, the Onit applications demo-ed in class — i.e., the backend where David and his staff configure workflows and dashboards — is surprising clean and simple.

David shows us the main dashboard he uses monitor the legal department (16 tiles of information).  He also shows one of the dashboards he built for Cam Findlay, ADM’s general counsel, which provides real-time information likely of interest and value to the C-suite.  Some of the tiles use Tableau to display the information graphically (other data visualization programs can be used).  All of these graphics are generated from data captured by Onit workflow systems.  The data are high quality because Cambria has ruthlessly reduced the number of unnatural acts required by his lawyers.

Diffusion theory wrap-up

Eric Elfman readily admits that Onit targeted Cambria as an early adopter and opinion leader.  Cf. Post 020 (discussing the crucial role of opinion leaders in accelerating innovation adoption).  Eric comments, “David got a whole lot of software for very little money. But we wanted him as a reference client.  And frankly, it’s been worth it.”

Cambria was drawn to Onit because it offered him the possibility of improving the performance of ADM’s legal department without requiring this lawyers to learn new technology or do data entry. This is the novel perspective of a true “visionary” customer as defined in Crossing the Chasm.

These are interesting anecdotes. However, if we want deep learning from this case study, it is important to tie what we see back to the empirically validated principles of diffusion theory.

As discussed in foundational posts 008 and 011, innovation adoption — whether it happens at all, and if so, at what rate — is primarily a function of five innovation attributes. See graphic to right.

In addition, software for managing complexity requires us to evaluate these attributes from two perspectives:

  1. Managers making the purchase decision. These are folks with a serious business problem and a limited amount of time and technical expertise, at least with software.
  2. Workers asked to use a new software solution. These are busy professionals who just want to get their work done.

Arguably, legal departments have historically made the mistake of focusing too much on (1) and underestimating (2). This explains their perennial disappointment with enterprise software.

The table below scores Onit from both perspectives using the simple scoring system developed in Post 011 (fast versus slow innovations):

  • Positive numbers (+1  to +3) speed up the adoption rate
  • Negative numbers (-1 to -3) slow it down
  • Mild effect = -1 or 1; moderately strong = -2 or 2; very strong = -3 or 3
  • No effect on rate of adoption = 0
Factor affecting adoption rate Manager Worker Adoption Analysis
Relative advantage 2 3 Managers get complete, high quality data, albeit after a learning curve. Workers are not asked to perform unnatural acts; minimal change management.
Compatibility -1 3 Managers are business process owners and have to learn cloud software related to workflow; new but surmountable. Workers get to stay within email and Internet browsers; basically this is change that feels like the status quo.
Lack of Complexity -1 3 Managers have to climb a learning curve, but its mostly cloud-based drag-and-drop tools. IT support is minimal. Workers carry on business as usual.
Trialability 2 2 Managers can get started at a low cost (e.g., just one Onit application) and build it out as needed. Worker feedback enables quick and inexpensive changes in process.
Observability 2 -2 Managers can see the high quality data pile up.  For workers, there is a limited ability to observe fellow knowledge workers being more productive. This factor is hard to change. It is also why we laugh at Dilbert cartoons.
Totals +4 +9

The key insight of this analysis is that Onit is likely to enjoy rapid adoption with workers, largely because it places so few demands on them.  Although managers don’t have it so good — they actually have to learn a new technology — it’s likely worth it.  As the ADM example shows, worker adoption occurs in a low friction way; also, senior personnel in the legal department can finally see, measure, and manage essential business processes. From a big picture perspective, this is a potential home run.

During class, Eric Elfman observed that technology start-ups are essentially “a series of experiments until something works or you run out of money.”  According to Cambria, Onit works well.  That is very good news for Elfman and Onit.