No One Asked Michelangelo About His Chisel

– David Wanis, October 2025

There are two seminal moments in an analysts’ career – the discovery of the DCF and the discovery of the shortcomings of the DCF. What feels initially like the secret to calculating intrinsic value to any business soon turns into the realisation the answer depends entirely on the assumptions the analyst has for the business, often many years into an uncertain future. Microsoft Excel made building a DCF easier, but it did little to improve the real-world accuracy of the resulting valuations.

In the mid-to-late 1990s and the early 2000s every Berkshire Hathaway AGM seemingly had a whole series of questions on how Warren Buffett calculated a valuation. What discount rate did he use? What terminal growth rate? What equity risk premium? How did he define cash flow, was it the same as EBITDA?

As if the answer to the investment track record of the greatest of all time was having the right number to plug into a spreadsheet.

Buffett and Charlie Munger had long suggested to their shareholders, when looking for their own investments they should focus on the quality of the business, which determines the sustainability of the returns and the ability to reinvest excess returns into future cash flow per share growth. If we look at the ~16 significant public market investments made by Berkshire Hathaway over the past 35 years, you would be forgiven for assuming the methodology was as simple as “find a great business and buy it for close to 10x pretax earnings”.

Source: Multiples are approximate pre-tax estimates derived from public analyses and financial reporting at the time of the investments. They can vary based on the specific timing and tranches of purchases.

The focus on the tool and not the underlying source of insight is like asking Michelangelo questions about his chisel. Here is the apocryphal equivalent of the Berkshire Hathaway AGM, from the 16th century:

“The Pope asked Michelangelo: ‘Tell me the secret of your genius. How did you create the statue of David – this masterpiece of all masterpieces?’ Michelangelo’s answer: ‘Quite simple. I removed everything that is not David.'”

At least Buffett and Munger were a little more forthcoming!

Many of the greatest businesses have their true source of advantage hidden. Some of the best research by investment analysts comes from identifying the true moat ahead of the market pricing it in. Domino’s Pizza Enterprises (DMP) is a useful case study for this.

From its listing in 2005, the market rewarded Domino’s for its superior operational execution. It strategically focused on a delivery-first model, using technology and logistical efficiency to outperform its then-dominant, dine-in-focused rival, Pizza Hut. This success was compounded by an aggressive international expansion strategy following the same playbook.

Source: Bloomberg Consensus Estimates.

I remember team stock discussions in the early days at a prior firm, when one analyst was convinced the Domino’s advantage was effectively ‘cheese arbitrage’ – the ability to buy cheese at scale and capture the retail spread. This led to some humorous images of the secret Domino’s cheese trading desk which they had kept hidden as their true source of profits.

In reality, the advantages the market focused on were technological leadership (hard to believe from a pizza store operator, but it was true), operational consistency at scale, and a strong brand and franchise model – although brand tends to be the outcome of a well-run business that consumers love and can fade.

It turns out that the true source of Domino’s moat was its quasi-monopoly on the ability to efficiently deliver food. For years, it was one of the few, most reliable options for fast food delivered to the home.

The scaling of third-party delivery aggregators like Uber Eats and DoorDash in the 2020s fundamentally disrupted this. Delivery logistics became a commoditized service available to any restaurant, erasing Domino’s key differentiator. Competition exploded and Dominos lost their direct-to-consumer relationship as the third-party platforms captured the aggregator position.

Source: Bloomberg, Company Reports

The share price of DMP has fallen significantly and the multiple has de-rated, but investors are arguably buying a different business in a different industry structure (with a very different – much more levered – balance sheet) than the history they observe and compare to. Maybe the turnaround works. Maybe Private Equity will take over the company. But in our view, the source of advantage has been seriously dented and the historic P/E multiple premium will be harder to restore.

Now that the “last mile” is commoditized, where will advantage move? This movement of advantage is not new by any means. Aggregation. Specialisation. Vertical integration. Horizontal integration. Companies and industries are in a perpetual state of movement as technology, demographics, regulation, global competition and innovation move the source of excess return.

One area where their power in the delivery channel allowed Domino’s to underinvest was the quality of the food. My personal opinion (with teenagers who still eat it): this is not good pizza.

In theory, one response to the commoditization of distribution is the improvement or differentiation of the product itself. Competing on product attributes or price rather than convenience. This appears to be the strategy of Guzman Y Gomez – build loyalty around the differentiation and quality of the offer, not convenience, speed or, if my experience has been representative, the quality of service!

This of course brings us to Artificial Intelligence (AI). Doesn’t everything these days?

Like a well-structured spreadsheet to do company modelling and valuations, or a research management system to capture qualitative insights and decision logs, AI tools allow investors to improve their investment process. But they are just tools – and while powerful, they are available to everyone so confer no more advantage than having a Microsoft Excel license or a Bloomberg Terminal. Once again it is how they are used.

At Longwave we are deeply interested in how technology can be used to improve the investment decision making process and performance outcomes for our clients. In the past 25 years there have been some incredible advances which have enabled significant gains across our business. In the investment world, tooling is often overlooked.

Ironically, asking Michelangelo about his chisel turns out to be quite interesting.  “He used a variety of steel chisels that he made himself, mainly a pointed subbia for initial work, a three-toothed gradina and a flattened curve unghietto (fingernail), and also a bow-driven drill (arco/archo).”

For our fundamental analysts, the objective is to reduce the amount of time required for necessary but lower value adding work and freeing up time for high value adding human specialist work – like figuring out the true and hidden source of competitive advantage of a business by talking to lots of people operating in an industry.

Once every decade or so, new technology comes along that the market decides is high-stakes, high-status and must be pursued with vigour. The Web back in the late 1990s. Big data in 2012/13 and of course AI today. The greatest tell in these times are a) people who don’t know what is possible with existing technology – so will try and get AI to solve problems far better suited to alternative, cheaper and proven technologies; and b) the need to “do something with AI” – we have this technology now, we have the budget, what can we force it into?

Steve Jobs provides guidance here:

“You’ve got to start with the customer experience and work backwards to the technology. You can’t start with the technology and try to figure out where you’re going to try and sell it.”

Yes, new technology allows you to solve new problems, but you need to start with the problem and work backwards to what solutions the new capabilities unlock.

News that OpenAI is paying bankers $US150 per hour to write prompts and build financial models demonstrates that time is indeed a circle. We will now have AI building DCF models which will allow an entire new generation of (artificial) analysts to make the same category of error all over again. And this time, they probably won’t have Warren Buffett and Charlie Munger to guide them to a more profitable path.