The 7 Pillars of Technology Breakthrough: Novozymes’ innovations in yeast show the new way

Over the years we have run 727 different stories around technology or commercial breakthroughs in The Digest — there’s even a venture capital group named Breakthrough Energy Ventures. Let’s explore today — using the recent news from Novozymes about their pioneering advances in yeast — what exactly a breakthrough is, and does. That’s relevant to everyone bringing forward a technology or product — whether yeast is in the mix or not.

The News

In Texas, Novozymes announced the launch of its next yeast technology, Innova Fit. Fit is the most advanced non-GM yeast in the market, our friends at Novozymes say  – eliminating production constraints caused by conventional and basic yeasts.

Fit provides ethanol plants with flexibility to achieve operational targets, while improving performance with enhanced reliability. The advanced Innova yeast allows producers to maximize inputs, achieve throughput, and production targets, without losing ethanol yield to common stressors such as high temperature. Moreover, Fit enables plants to significantly improve throughput and plant efficiency.

What’s important?

The key word we’d offer is one right out of the Burger King playbook, “have it your way,” which is to say that what Novozymes is accomplishing in the yeast business is to drive it farther and faster into the Customization Age.

It’s something that Lallemand recognized some time ago and they pushed through into an age of yeast customization, and away from the one size fits all era. With Novozymes putting their considerable shoulders to tasking in a similar push towards customization, it is well worth shining a light on.

As Novozymes Brian Brazeau expressed it to me, “What is the unmet need? We want to provide customers with an ability to run their plants their way. They are looking to get as much profit as they can, and we want to allow them the flexibility. Here what we are bringing to the market is a robust yeast that is reliable and did what we said it would do. It’s a very competitive space and we should be very humble about it, but we think we have a great start.”

Let’s start with a truism that is no less true simply because it is widely understood. Yeast has different economics in Iowa than one on the West Coast, and operators would like to have their choices enabled by their yeasts, not constrained by them. For example, in the summer time we have high temperatures, and we have seen instances where there are carbohydrates left over and not converted, unless there’s a customized yeast that can push through summer conditions.

But there’s more to the custom need than just the seasons. Some plants prefer longer fermentation runs, and it seems obvious that a yeast that can better tolerate higher alcohol levels for longer periods will have more value. And there are corresponding special needs for shorter fermentation runs. And one day, we might see that certain yeasts will work better with certain enzyme cocktails — which are highly customized in their own right. So, customization has opportunities at the system level.

“We think the solution is flexibility,” said Brazeau. And as Novozymes attracts attention for its FIT release, let’s mark customization down as one of the primary drivers. A toolbox is coming, a toolbox of yeasts, like a toolbox of screwdrivers or drill-bits.

Where does customization sit in the development cycle, why does it happen now, not sooner or later?

The problem that we may have is simply in the way we think about product adoption. Typically, we use a bell curve to describe the transition from a base of pioneer customers, to early adopters, then mainstream adopters, late adopters and finally laggards. It looks like this:

It’s a good way of understanding, when looking backwards at an innovation, what happened in terms of adoption, but it really doesn’t look to the forward question of “why do breakthroughs happen?” and “do breakthroughs have cycles?“. It’s an important pair of questions, because product stories do not end with 100 percent adoption and a joy of a sale to the laggards, they finish up with an “end of life” notice and a discontinuation, amidst new waves of product releases. So, what happens?

Bell curve enthusiasts offer this curve, as an alternative.

We might ask, though, what is driving the growth? What is driving the decline? This bell curve helps us with the what, but not the how or why. As a technology developer and observer, over many years, I have found it more useful to think in terms of 7 Pillars of Technology Breakthrough, defining a breakthrough from the customer point of view as:

“A material improvement in performance that is worth the adoption risk.”

I like that definition because, inherently, as the product is rolled out, the risks become smaller, and therefore the material improvements have to be more “unobtainable any other affordable way” in the earliest moments when the adoption risks are highest. Later on, as risks diminish, the barriers to adoption become smaller. ‘Performance’ could represent, “the same yields as everyone else, at a lower price,” or “higher yields than anyone else” or “a novel application”, and so forth.

Here, we have a method that explains why we have this expansive period of growth as we reach that “growth market stage”. The risks are falling, quickly, relative to the improvement in performance. Later, the material improvements diminish — and more importantly, there are some stubborn risks that always remain even when technology risk has vanished (e.g. agricultural commodity risk, weather risk etc.) — eventually, risks are falling relative to performance — only, more slowly. Eventually, improvement stalls, competition arrives in the form of same-as or much-better products, and the product commences its journey towards end-of-life.

Though risks will always abound, Currency risk, technology risk, country risk, finance risk, disruption risk, company risk, supply chain risk and so forth.

The R&D period can be neatly stuffed mostly into the “vital breakthrough in prospect”, where the partners collaborate to bring forward a desired improvement in performance, sharing the considerable risks through the risk-sharing that partnership is grounded in.

This explains the nature of partnership and collaboration. The R&D stage is a stage where the risks are so high to an individual company that the material improvement in performance is simply not worth the risk — hence, we spread the risk. In investing, we spread risk with a portfolio of investments. In technology, we spread the risk with a portfolio of investors.

The more risk, the more partners, and ultimately the need for more non-dilutive equity, the kind that government investment brings, because it spreads the risk without dilution of the rewards to the developers, and it also gives more assurance to customers that a proposed technology may, in fact, reach the market and therefore is worth devoting customer resources to testing, validation, and strategic investment.

The Breakthrough Stage

The Breakthrough stage is relatively easy to understand. At this point, the technology has validated its material improvement in performance, the risks have diminished to the point that the product can move from development to deployment, and production soars either as a small group of companies deploy widely, or a large group of companies deploy narrowly. During this period, we usually have one-size-fits-all limitations, and optimization has hardly begun because of the allure of those “material improvements in performance” that one has with release 1.0.

Frankly, you could hardly give away a 2006 iPhone today, but in 2006, people took sleeping bags to stores and camped out in vast lines to become early adopters. Optimization would come later, customization would come later. In September 1929, as the Jazz Age ended, F. Scott Fitzgerald wrote glumly to his book editor Maxwell Perkins: “The Post now pay the old whore $4000 a screw. But now it’s because she’s mastered the 40 positions — in her youth one was enough.”

The Era of Optimization

The opportunities in optimization are so profound, and there are so many players seeking them, that whole companies with enormous valuations such as Zymergen, live more or less inside the Undiscovered Country of performance enhancement. Gene editing, machine learning, deep learning are all saddled up in service of Optimization. Here we are learning from data, in order the maximizer the performance of a given machine or process in a given task. A task could be, for example, to make more product from a ton of biomass, or to make less of a given by-product, to make an organism more tolerant of stress, or more robust in the face of toxic conditions, or to conduct a given process at a lower temperature or pressure, or to tolerate higher or lower acidity in the water. Just to name a few optimizations out of hundreds.

We’d hardly call this is “Optimization stage”, because it can go on and on and on, into eras, or eons. We are still optimizing yeast fermentation into wine many thousands of years since Noah trampled his grapes. We are optimizing beer production, fuel production, crop yield, crop protection results — not just through new products and process but in making the old ones better. I could be working on rate, titer or yield.

Perhaps this is the most crucial stage in determining the useful life span of a product — the more we optimize, the longer it lasts in the face of competing processes or products. The petroleum business would not exist anymore had it not been greatly optimized since the 1860s. Then, we threw the light cuts away — gasoline, that is. There were many crudes that were almost impossible to handle. We didn’t make as much gasoline as we wanted to. The more we optimize, the more that some other process earlier in its life cycle will struggle to offer that “material improvement in performance that is worth the adoption risk.”

And, optimization occurs at several levels. There is the organism, the catalyst, the sub-process, the process, the unit, the system, the site, and the sector. In most cases, sophisticated players work on several levels at once. Sometimes with in-house staff, often with vendors and partners, from the smallest level or organism optimization where consortia might work together, up to sector optimization when industry associations might foster collaborative work ranging from standards to regulations and policies.

The Customization Epoch

We’ve been through the customization story in the opening sections of our story today, since that is stage that yeast has apparently reached. But let’s spend a moment on value, which is to say that the Age of Customization can be the period of maximized value to the technologist, because the value of the underlying technology is being maximized for every customer. It pens up opportunities to gain new customers for whom the one size fits all era did not provide that “material improvement in performance that is worth the adoption risk.” And these new customers, and existing ones, come in at maximized values.

As you note, Customization is a special class of Value Optimization, in which we are maximizing the utility to the customer. And it should be the Era in which the prices are the highest because the value is the highest. In short, it’s the High Point of a product’s value cycle.

The Stage of Price Competition

At some stage, new products emerge or new competitors find ways to create the same outcomes without treading a competitors’ patents. Absent collusion or cartels, the Era of Price Competition begins — something that transfers value to the customer and away from the technologist in the form of lower prices. It does not necessarily follow that lower margins result from lower prices. Because, Technology companies typically respond by investing in Economies of Scale and they also transfer the impact of falling prices backwards through the supply chain. In short, when ethanol prices fall, corn prices fall, and ethanol plants consolidate, to give an example. The Stage of Price Competition can continue for an awfully long time — it is the era which more or less the petroleum refining industry is in, building more and more capacity but fewer and fewer refineries, and transferring any pressure on fuel prices right back into oil prices, or all the way to the beginning of the value chain. Refineries are getting bigger even as they are getting more optimized and customized. It is the future of every industry, and the bioeconomy’s future too, though not here except in a few select cases, such as corn wet milling.

We are still seeing “a material improvement in performance that is worth the adoption risk,” and by this stage the adoption risks are quite low, and thereby the material improvements can be pretty slight in order to justify adoption. Often in these sort of environments, we can see quite a bit of customer switching, as we see with mobile phone plans today. The switching is in part a response to shifting prices and price competition — but it is also the result of that falling adoption risk that makes price takers of us all.

The Stage of Margin Compression, and the End of Life

At some point, price competition is continuing, but economies of scale are harder to find, supply-chain optimization efforts falter, and new products keep coming in all the time. We have entered the period of Margin Compression, where companies will tolerate legacy products and processes because there’s still sufficient margin in them. At first, the margins are higher, perhaps, than for other proposed investments, and there may continue to be reinvestment in the product in the form of optimization efforts, customer acquisition and so forth. Eventually, the margins fall below the corporate hurdle rate for re-investment, which means that the technology is maintained but no longer improved and investments in sales force begin to falter. Margins begin to compress faster as reinvestment slows, and a runaway reaction begins, with margin compression causing the withdrawal of resources, causing more margin compression, and so on.

Until we can no longer find“a material improvement in performance that is worth the adoption risk.”

Because, at the very end of life, adoption risk begins to rise, the risk of being left adrift with a cancelled or unsupported product. You can buy a Mac G3, once a breakthrough product, for practically nothing, because modern software doesn’t work on it, and what felt like blazing speed now feels like a crawl.  Customers begin to stay away from a product for many of the same reasons that young people tend not to marry very old people no matter how charming that older person might be.

And so, that brings us to the final stage, which is the decision to discontinue. Sometimes, these periods can take a long time, owing to contracts with key customers or Y2K dependencies that cause us to keep one process alive because the disappearance of one (unviable) process, may cause run-on impacts in some other (viable) product or process. As an example, almost all of the technologies we see at Colonial Williamsburg have long since gone away, but we preserve them at Colonial Williamsburg as a demonstration of the 1770s-era way of life, and Colonial Williamsburg is a going concern and needs these technologies to be kept alive even if only in the most limited way. Coopers, tanners, blacksmiths and so forth — technologies gone by, but still not End of Life because of living museums.

The Bottom Line

Where do we make money? All along the life cycle, but in different ways at different times. In many ways, industrial biotechnology differs little from the Bell Curve description of the life cycle. Differences few in number, however, are great in importance. Most vitally, a more detailed and form-fitting model — more Optimized and Customized — if you will, performs the same function that Strain Characterization plays in strain engineering. First we have to discover and analyze what’s already with us in the living world, before we can begin to think and design improvements to our system.

What can be improved in the lifecycle? Most of the vital improvements with the most impact, like the Big Bang, happen in the earliest stages of technology formation. We have lots of directed evolution of microbes but less directed evolution of lab work. Most of our innovation heads down this path:

Idea, technology, application, proof of concept, company formation. And one of the Digesterati once noted to me that the return of research grant investments at universities are appallingly low. Forget meeting corporate hurdle rates like a 30 percent return in 3 years. Think a 90 percent loss over 3 years and you might be closer to the mark, so I was told by one in a position to know. I don’t believe for a minute that such a result comes from bad minds working on useless problems. Rather, the venture is being introduced too late into the invention cycle — rather, as the result of an invention.

The Edison Company accomplished much because it was directing invention in the 19th century, rather than in adopting winners burbling up from the labs. Tesla was directing his inventions, and thinking in terms of a system of innovations, rather than an innovation. It’s easier to see the value, in chess, of investing in a Knight, if the other pieces are already imagined, or imagined at the same time. Invention can come from targeting and in parallel, not just the serial fashion in which we usually innovate. Finistere Ventures has been working in recent years more along these lines, and it’s welcome to see.

Perhaps one day more of technology development will follow the Edison path. It’s how we proceeded to break through on atomic weapons and energy, the very technology set which led to the formation of our US National Lab system. The concept of atomic power wasn’t really proved, at any scale except for projections from testing of tiny units that had not been tested in an integrated form, until the Trinity test in July 1945. Industrialization preceded certainty and parallel development preceded an integrated demonstration. It was also how we landed on the Moon, more or less, despite launching two Apollo missions before the Lunar Module had ever been fully completed or tested in space.

We do build systems of technology this way, and there are risks. (Anyone can whisper “cellulosic ethanol” at this point — a technology that was developed on short timelines and requiring multiple breakthroughs in parallel.) However, the deployment companies were not formed at the time the technology was born; rather, the technical advisory on the Renewable Fuel Standard targets and timelines came primarily, according to a massive report we read  by Hanna Breetz on the RFS’s birth, from technologists and backers of those technologists. As we understood it, the nascent companies in the space expressed much greater caution on the timelines that in fact were enshrined in RFS targets. Had companies and extensive partnerships been formed earlier, with robust financing from a spreading of the risk beyond venture participants, we may have seen a better result even from that storied chapter in the bioeconomy’s history.