Blockchain as an Interesting Waste of Time: How Blockchain’s Dismal Prospects in the Voting World Translate to the Enterprise.

I should start off by saying I’m firmly in the blockchain skeptics camp, lurking somewhere between the luddite and nihilist wings of the anti-blockchain gang. I’ve seen too many new and interesting technologies become obnoxiously over-hyped in my career to believe a tenth of the ridiculous claims about blockchain’s ability to solve every major problem in business and civic life.

That said, my disdain for blockchain until recently was mostly fueled by the fact that too many of its proponents seemed to be Dr. Oz-like in the hyperbolic nature of their claims, and, like Dr. Oz, were unable to see the silliness in some of their claims.  Meanwhile, the stench from the ongoing bitcoin ICO scam/ponzi scheme continues to taint even the most serious of blockchain endeavors.

So with that not-so-well-hidden disdain as background, I recently attended an analyst event held by SAP that highlighted the commercial prospects for blockchain, an event that included some interesting customers talking about some very interesting projects involving blockchain proofs of concept. And the result of that nth briefing by the nth vendor about the prospects for blockchain in the enterprise finally helped me crystallize my skepticism into a single sentence: so far blockchain is trying to either solve problems that aren’t real problems or solve problems that already have viable, albeit mundane and “unsexy,” solutions.

In other words, I think blockchain is at best an interesting waste of time and, so far, not much else. And at worst it’s a distraction that could damage real efforts to do a better job at solving problems that really matter. The reason boils down to this: in every case I’ve examined thus far, if you really look hard at the ultimate goal of a blockchain POC – and we’re not talking playing around with blockchain enough to be able to brag to your investors, partners, or next boss about how leading-edge you are — there are better, more effective, faster, and cheaper ways to improve the broken and/or inefficient process that is being blockchained. Much better and more effective than taking a flyer on a new, largely untried, and overly hyped concept.

To explain this admittedly contrarian viewpoint, bear with me as we journey over to the world of voting, where a number of individuals and organizations are claiming that blockchain is the answer to the an election security problem that is suddenly and unfortunately part of the voting landscape of our country. (Why voting: Some of you know that in my copious spare time I serve as the volunteer CTO of US Vote Foundation, a non-profit, non-partisan civic tech organization. Shameless plug alert.)

First, let’s get grounded in what the ultimate goal of everything we do in voting should be about:  increasing the number of valid votes cast in every election. This is an important nuance in a domain where, to its proponents, blockchain is being touted as a way to allow electronic voting to finally take place in a secure, safe manner. Importantly, while they are often conflated, these problems are not one and the same thing.

This causation – blockchain solves the security problem that prevents voting from being a one-click experience on your smart phone and therefore dramatically boosts voter participation rates – is well-meaning but problematic. And emblematic of much of the magical thinking around blockchain in the enterprise. There’s no doubt we desperately need more voter participation in order to preserve the integrity of our democracy: it’s a given that if “did not vote” was a candidate in a general election that candidate would win every time. Locally, it’s even worse: local election participation rates are often in single digits, allowing special interests and corrupt individuals to do serious damage to local institutions and budgets.

While there are people who think that the solution to low voter turnout is to make it as easy as Amazon one-click, that sad truth is that so far there’s simply no evidence to support that assertion, and most apps and much of the work that groups like Google have done to create more app-like voting experiences simply haven’t moved the needle very far, based on what we saw in 2016 and the ensuing elections. Applying an even more arcane and complex tech platform like blockchain to this problem doesn’t augur well for a solution.

Okay. So at least blockchain will solve our nasty voting security problem, right? Unfortunately that too is another red herring. Voting security is much less broken than the hypesters would lead us to believe – there is zero evidence that anyone was able to actually change votes and sway the election in 2016 or any time since. (The NYT’s overview article on the case for Russian interference in 2016 makes no mention of hacking actual votes because there was no evidence that actually happened. The 2016 election hack was about creating undue influence, not stealing votes.)

Not that it isn’t possible – some recent over-hyped events have posited that a 11 year old can hack a voting machine, which I don’t dispute, though there’s are some serious questions about whether what is possible is actually probable and could be done at a scale big enough to hack a national election.

But all of this talk is a distraction from the real problem, which is increasing voter turnout. In the real world of U.S. elections, unconstitutional voter suppression and gerrymandering (look at this map of Birmingham, AL – those red lines are the voting district boundaries of the city! Guess what the race is of the majority of the people who live inside the redline versus those who live outside.) are much greater and much more real threats to voter turnout, and maximizing turnout is of course what we’re really talking about here. We do have the problem of hanging chads and other mixups in vote tabulation, and to a certain extent better technology could solve this problem. But so could a better system of more standardized paper ballots (so say lots of election security experts as well as the National Academy of Sciences). More systematic verification and audit processes could help too. These are known solutions that work, and in my opinion there’s so much room for improvement with what we know works that the temptation to leapfrog into the unknown should be set aside until these easy wins have been implemented. Keep this in mind when we segue back to the question of blockchain in the enterprise.

 

Another consideration for which voting is a great proxy for enterprise blockchain is the whole issue of change management. Blockchain by definition will disrupt and/or replace existing processes, whether in the business world or the election world, and assuming that the change management component of that process replacement will be simple  and easy to realize is to be in serious denial about the impact of this degree of change.

Change management in the election world would be as hard or harder than it is in the enterprise, and  that means we need to very carefully consider the cost and complexity of sending thousands of election officials across the country to blockchain voting system training, and, most important of all, providing them the budgets to actually implement this new (and, I repeat, untried and unproven technology).

Finally, even if there was some magic bullet to deal with the change management issue, some seriously smart people have serious reservations about whether blockchain could work for voting, particularly because it would require secondary systems to handle the unique privacy and verification requirements of voting. As the article in the link above explains, a blockchain voting system would need to do something about the inviolability of the public address/private key design inherent in blockchain technology: as a voter your private key would need to be accessible to potentially multiple election officials in order to do an audit, meaning a secondary “key” system would have to added for voting that wouldn’t be so private. This breaks the security model of blockchain and the intent of the overall concept. If multiple election officials can verify a voter’s identity and the vote she cast, the same credential can also be used to change or invalidate a vote. Considering such a blockchain voting system would centralize a vote counting process that is almost pathologically decentralized today, the addition of the secondary key would actually make voting more vulnerable, as it would provide a single point of entry to a greater number of votes than might otherwise be available to a hacker today.

In other words, it’s a real possibility this might not work. And instead of wasting time on a dubious venture, the election world might be better served by putting blockchain on ice and focusing on a host of proven, simple, and low-cost solutions that can actually help to improve turnout and not try to solve a technology problem that exists more in theory than in reality.

In case you didn’t see that as a warmup to my larger point about blockchain in the enterprise, let me spell it out clearly: pretty much every enterprise POC using blockchain is trying to solve a problem that could be solved better and more cheaply – and possibly more broadly –than by using a blockchain-based solution. And while well-intentioned, and downright fun, these POCs are drawing the best talent in the enterprise to solve real problems with what is for now a marginal, unproven technology. That talent could and should, in my opinion, be put to work on implementing the solutions that actually work today: they may be much much less sexy than blockchain, but the impacts will be orders of magnitude greater than otherwise.

Let’s take blockchain in the supply chain. There were a number of great examples of blockchain POCs in the supply and logistics chain at the SAP event earlier this month. Track and trace is a major issue in supply chains, and the inability to understand whether a particular component or finished product is the real deal and is in the right place at the right time are huge problems that touch profits, customer satisfaction, and regulatory compliance, to name just a few of the impacts.

Naturally track and trace is something blockchain could do a great job with – an unbreakable blockchain could, at least in theory, ensure that those brand name jeans or that next-gen microprocessor or that high-end fish filet is what it’s supposed to be and not something it’s not. We can agree on that, no problem. In theory.

But there are already a lot of solutions available – test, tried, and true – that can solve many of these problems, if only they could be adopted as broadly as possible. Take the logistics chain – blockchain could make sure that a pallet leaving the factory in Vietnam was the same one that arrived at Walmart with a significant degree of certainty (assuming that there was no way to actually tamper with the pallet’s contents, a major assumption IMO. More on this below). Great idea, except that Infor’s GT Nexus has been solving this problem for years by tracking the existing paper and electronic trail of supplies and finished goods as they wend their way from the manufacturer’s loading dock to the retailers loading dock. Other vendors do this too, many others. In a variety of ways, most of which work wonderfully.

Fundamentally, the reason we still have a problem with provenance in supply chains is more about the fact that the existing solutions haven’t been adopted in a sufficiently broad manner and companies across different vertical supply chains haven’t updated their technologies and business processes to fine tune this problem. I’m reminded of this every time I look at the packing slip that came with the product I recently ordered from AceHardware.com and picked up at my local store. It looked like this:

This is, of course, a printout from an AS/400 or similar system that Ace still uses in its warehouses. So does a major outdoor equipment vendor I dealt with recently. (Excerpt from that exchange: “could someone check the AS/400 system to see if Mr. Greenbaum’s pack is still in the warehouse?”). Does anyone think that either of these companies should spend time on a blockchain POC to make sure my new backpack or my 40lbs of pool salt have left the warehouse or not? I think they need to move some of their back office technology into the early 21^st century first. Once that’s done, let’s talk about getting some of that blockchain stuff going.

Even in a regulated industry like pharmaceutical manufacturing, vendors like Kinaxis are able to use existing technology and data to significantly move the needle in solving a range of problems in the supply chain. And along the way they’re showing that the biggest single hurdle to supply chain success, after change management, as in getting people to work and act differently, is data quality. Fix that – not a trivial problem, granted – and a massive host of problems are on the way to being much much better managed, including the provenance of  the supplies and products that help save patients’ lives. Again, you can fitter away a lot of effort on a blockchain POC or you can start improving the data quality in your supply chain now. The latter is a known problem with a known solution, and for my money that’s the better way to go.

Meanwhile, in supply chain/blockchain POC land, there’s the problem of actually verifying at the point of consumption that what’s on the pallet or in the container or in the pill bottle is actually what was inside when the item started its journey to the consumer. If I do pallet-level tracking, why couldn’t someone just remove the items from the pallet and replace them with counterfeits? I can tag a fish filet at its point of origin (which is what one of the POCs at the SAP was all about) but does that really ensure that the diner isn’t going to get turbot instead of halibut? Blockchain notwithstanding, thievery as the second or third oldest profession has a long history of hoodwinking even the most technically adroit solutions. I’m pretty sure a smart hacker – okay, maybe one who is a little more than 11 years old – could subvert pretty much every blockchain POC I’ve ever run across.

All this is to say that sometimes we’re tempted to find a problem to solve that isn’t really the one that will make a difference, and in trying to solve it we risk diverting talent and money in the wrong direction. I don’t think anyone should stop experimenting with blockchain, in voting or supply chains or anywhere else – experiments are the only way to test hypotheses and move human knowledge forward. But we should stop and ask whether there are bigger problems or more focused problems that need solving before we buy in too much and over-commit. Right now, blockchain deserves experimentation, but pretty much any domain where there’s a blockchain POC underway probably has an easy half-dozen tried and true solutions ready to take up the call. Let’s at least try hyping those first before we get too far down the road with blockchain, and waste too many precious resources solving problems that have already been solved. Blockchain’s time may still come, but today is not the day.