The ability to estimate numerical quantities is at the core of what being a physicist is all about. Ian Taylor explains why that skill is also handy in the business world to decide which technologies are most likely to succeed
Physicists (and engineers) are great at estimating. We’ve obviously never counted the number of electrons in the visible universe, but we reckon there are between 1080 and 1082 of them. As for the “Planck length” at which quantum effects start to dominate gravity, that’s estimated to be 10–35 metres. The transition between laminar and turbulent flow in fluids, meanwhile, comes in at Reynolds’ numbers of between 2500 and 4000.
These estimation skills, which are a familiar part of any physicist’s toolkit, can also come in handy for anyone working in the business world. In particular, the ability to estimate is useful when examining the business cases drawn up by start-up companies seeking to drum up financial support from investors and banks.
Most business cases completely overestimate a company’s predicted profits and make wild – and mostly nonsensical – statements.
Such documents inevitably paint a rosy picture of a firm’s potential because, without money, nothing can be done. However, in my experience, most business cases completely overestimate a company’s predicted profits. They also make wild – and mostly nonsensical – statements about the transformational impact it will have on a particular industry.
Now I appreciate that business cases need to contain a bit of hyperbole if they are to attract investor funding; there’s no point being unnecessarily gloomy and down-beat. But given that up to 90% of start-up companies fail, it’s essential to have the ability and skills to evaluate a start-up company’s business case as realistically as possible.
Here’s the rub
Let me give you a specific example from tribology – the science of friction, lubrication and wear. Tribology might seem mundane, but it is a big deal, especially when you realize that 25% of the world’s energy is used simply to overcome friction and wear. (The Institute of Physics has a special interest group in tribology if you want to find out more about the field.)
One topic of huge interest, on which Google Scholar reckons almost 15,000 papers were published in 2025 alone, is that of nanoparticles as potential lubricant additives. Despite that large amount of research, I am not aware of any mainstream commercial lubricants that use any of these additives. So why is there such a large disconnect between academia and actual practice?
This is where our estimation skills come in. We know that lubricants are mostly hydrocarbon oils plus chemical additives such as antioxidants, friction modifiers, dispersants, detergents and anti-wear and anti-foaming agents. According to data from Kline and Company, about 40 million tonnes of lubricants are used annually around the world, roughly evenly split between industry and the automotive sector.
Automotive lubricants typically contain 15% additives, whilst industrial lubricants have about 1%. The “average” additive content, over all lubricants, is therefore about 8%, which equates to 3.2 million tonnes. With Polaris Market Research estimating that the lubricant-additives business being worth about $20 billion, we can see that the “average” additive price is roughly $6250 per tonne.
Particularly important to lubricants are anti-wear additives, which stop metal surfaces from breaking down when the oil that normally keeps them apart has disappeared. Lubricants usually have about 1% of such additives. Given the numbers above, we can estimate that the overall market for them should be about 400,000 tonnes.
Can entrepreneurship be taught? An engineer’s viewpoint
That’s not far off the real market size, which business analysts Markets and Markets reckon to be about 300,000 tonnes. (The figure is slightly less than my estimate because some lubricants, such as transformer and circulating oils, don’t contain such additives.) With that market being worth almost $1 billion a year, the average price of anti-wear additives is about $3300 per tonne.
There’s also another key component of modern engine oils, namely friction modifiers – “slippery” substances that form films on moving surfaces to reduce friction and stop them from shearing. Some of the best friction modifiers, which are based on molybdenum, cost about $10,000 per tonne. Given that base oils typically cost between $1000 and $2500 per tonne, molybdenum-based friction modifiers are much pricer than anti-wear additives.
Graphene isn’t gold
Armed with this information, we can now see why graphene – the “poster child” of lubricant nanoparticles – has flopped so badly from a commercial point of view as a lubricant additive. This 2D carbon material costs a staggering $50,000–$200,000 per tonne, which is more than five times pricier than the most expensive additives in lubricant formulations.
The message is clear. For lubricant nanoparticles to become commercially competitive additives, their price per tonne needs to be come down by at least an order of magnitude. Now most Physics World readers are unlikely to be in the business of selling lubricants. But for anyone developing a new technology, the point is you cannot assume customers will come running.
Businesses need to be aware of what existing technologies are out there – and how much those rival technologies cost.
Businesses need to be aware of what existing technologies are out there – and how much those rival technologies cost. If a new technology is too expensive for that application, different markets may need to be considered. Graphene, for example, has proved more useful in composites and materials reinforcement, coatings, energy storage and electronics, rather than lubricants.
Looking at new technologies even more broadly, the most successful by far are those that have fallen in price year-on-year for a decade or more. In computing, for example the cost of CPU time per dollar has typically dropped by a third every single year for the last 70 years, while computer memory has plummeted from $604trillion per terabyte in 1959 to just over $1000 per terabyte in 2023. As for electric cars, they are selling like hot cakes now that batteries have dropped from $10,000 per kWh in 1991 to barely $100 per kWh today.
My message is simple. When it comes to any technology, a sure-fire way to judge if it’ll succeed is to see if its price is likely to drop. No-one can know what the future will hold, but as physicists and engineers, we have a great ability to make estimations. By applying those skills when we sift through business cases, we’re more likely than most to identify the real gems from the rest.
