Linoleum Floors & Fluorescent Lights: Where Your Paper Records Go To Die
By Peter Hébert
At Lux, we’ve long been believers that any job that is dirty, dangerous and dull is best done by robots. Intelligent machines (aka robots) have been a big Lux investment theme over the past several years — with well north of $100M invested.
Dirty, dangerous, dull….but what about disturbingly illogical? Some of our highest conviction investments followed after our team shouts out in dumbfounded unison a simple rhetorical question. It’s a question elicited after hearing some industry process, that people take for granted, explained to us in unemotional, exacting detail: “Wait, we currently do WHAT?!”
We’re excited to announce today that Lux is backing Ripcord, Inc., a venture to make digital sense out of physical chaos. Together we are taking on the vast, sprawling (literally), $25B records management industry and all the archaic archives that got their start in the Atomic Age burying files in depleted limestone mines to protect from nuclear disaster (*True story). By 2018, Ripcord is projecting a digitization rate of 50M records/day, a rate that is orders of magnitude greater than legacy solutions. Founder/CEO Alex Fielding is a rare entrepreneur. We’re also working with other like-minded investment partners to prepare for the journey — KPCB, Legend Star and Apple co-founder Steve Wozniak.
Dunder Mifflin be damned. For years, we’ve been talking about the paperless office….and…wait for it….umm…it’s still just a dream, yet to be realized. While more and more communications (text, video, email, messenger) make electronic hops and avoid print altogether, there’s still a big, fat (growing!!) load of paper that is birthed into existence around the world every day. Paper records are stubbornly clinging to life alongside new tools like DocuSign and Slack, and not-so-new tools like email and document management systems like SharePoint and Box.
Let’s be clear: paper records contain vital information for complex companies. Business records are an asset. Records contain immense historical context, reams of training data ripe for analytics and machine learning algorithms. But it’s as if an incredibly valuable, deep knowledge base lies right before our eyes, except for the fact it is unreadable (at least digitally). Quite simply, the technology to deliver fast, cheap, accurate digitization has simply not existed.
In their boxed and dust-gathering dead-pulp paper form, records are a useless liability. And a costly one at that. Iron Mountain reaped more than $3.5B (*with a B) in 2016 revenue for the privilege of providing Global 1000 companies’ paper offspring with cozy shelf space, tucked away in bankers boxes to sleep under flickering fluorescent moonlight.
Nothing against Iron Mountain (in fact, it’s one hell of a business — generating $1B in adjusted EBITDA last year alone), but they are like car impounds for paper information. Companies and their people produce valuable intelligence and information. But then they elect to pay Iron Mountain (and many others of their ilk) to store them away in vast underground vaults, with virtually no idea what is written on those pages and no way to access them. When companies want access to these records, they have to pay separate — exorbitant fees — to have the documents shipped back to HQ. It costs companies too much money and takes too much time this way.
Why hasn’t this industry been “disrupted” yet? Two reasons:
1) Misaligned incentives in the market. The largest incumbent in the space, Iron Mountain, is a public REIT. As a REIT, it is measured on the market by its tenant occupancy rates. If you think of Iron Mountain as a landlord, then the boxes of records they store are their “tenants.” Then Iron Mountain is incented to keep these boxes on the rent roll. A box of records that is digitized is a box that is no longer a tenant. So digitization threatens to cannibalize their core business.
2) Digitization is hard. Contents of these bankers’ boxes of records are highly variable — different types of records in boxes, different orientations and sizes, etc. Ripcord has an expanded definition of digitization. It goes beyond just scanning. It also includes using AI to make records useable searchable, manageable with as little human interaction as possible. Still need people to train the system, but the combination of using humans and AI is critical for accuracy/data integrity. This requires a multidisciplinary approach that involves robotics, machine vision, mechanical engineering, embedded sensors, industrial engineering, operations management, machine learning, artificial intelligence, and data science. Ripcord’s engineering team is uniquely qualified to do this.
Congrats to the entire Ripcord team. They’ve spent two years of hard work getting to this point. What they’ve achieved is impressive. So now the hard question: what will we do with all of those depleted limestone caves? Bowling alleys?