Valuation crutches

For the past few years, it has been all too common to see a pre-profits, and sometimes pre-revenue, investment proposition that is valued entirely as a multiple of revenues. Sometimes it is valued even more narrowly on annual recurring revenues – past, present or future. To my mind this is not far off valuations done on the back of a fag packet.

The valuation may be supported by references to comparables of highly variable relevance. You see start-ups bragging of their unique value-added proposition, but still being very content to use the buoyant revenue multiple valuations of (supposed) non-competitors to value themselves.

Extreme venture capital cases value themselves by taking forecast year-five revenues and applying a chosen revenue multiple to compute a ‘terminal’ value in year five, which is then discounted to a present value per share. It has to be said that five-year forecasts are rarely ever accurate – nearly all miss significantly, and almost always on the downside. Consulting an astrologer would be no less useful than this sort of computation, but the placebo effect of a horoscope generally comes at a much lower cost.

The revenue model perhaps best fits businesses with long-term contracts. Clearly, if you have a highly profitable business that gets 10-year fixed term contracts, then revenue multiple valuations have at least one foot on the floor. But such contracts are very unusual, although highly desired by private equity investors. Much more common are one-year contracts that may be renewed – or may not. ‘Recurring’ is then a judgement.

The first variable to look at is the cost of acquisition of customers. This is the area where the most common venture capital ‘financial engineering’ happens. Spend, say, £10,000 to acquire a customer’s 12-month contract with revenues of £5,000 pa; now multiply the £5,000 by six (because that’s the ‘industry revenue multiple’) and you have created £30,000 of company value. Even better, capitalise the acquisition cost rather than expensing it so the profit and loss looks more cheerful, too.

The most extreme tricks have gone so far as to raise equity and then to lend some of that to ‘friends and family’. This money is then used to pay for contracts – the company’s value increases mechanically, and you raise more equity at even higher prices. That pyramidal trick can be repeated on a rapidly growing scale. 

Of course, revenues are only valuable if they generate profits. Low gross margin sales generate less cash flow and actual value.

Customers are key

Then there is customer stickiness. If customers seem likely to stay for ever, then these contracts attract more value than a business model where contracts are rarely renewed. Competitive pressures exist in virtually all markets where the rule is that price pressure and technological change will likely greatly reduce the future value of a revenue stream. 

Mostly you can only guess at these impacts. Not so many years back private equity was killing to buy Yellow Pages businesses; the internet has greatly eroded the value put on that.

Revenue multiple valuations only work if sales and marketing costs and productivity, revenue retention and product profitability are all accurately assessed. That, too, may be essentially impossible to do.

If a company is in a business segment where there are multiple similar organisations, then its market valuation will likely take some account of all these factors. But the public markets can be woeful at valuation. Witness the (luckily brief) over-exuberance of shell companies in the US – SPACs (special purpose acquisition companies) – where sentiment overpowered the use of intelligent analysis. Revenue multiples were widely used to justify the improbable.

You clearly don’t need to be a financial genius to be very wary of revenue multiple valuations. Their only clear benefit is their ease of production and – as so often the case – idleness is a poor business strategy. 

Successive UK governments’ plans to create a smoker-free society may mean the next generation will not have any fag packets to make computations on.