Rightmove’s violent share price movements following its trading & technology update and analyst meeting last week were a reminder of how important it is to think about the value chain and business model rather than assume that tech and tech change is what matters most.

This is particularly important in addressing the bubble watcher’s question of whether in many industries short-term ‘soft’ AI gains in areas like marketing can be sustained before long-term ‘hard’ AI gains in such areas as manufacturing or development are delivered.

Rightmove’s decision to accelerate its AI spend and to take a short-term hit to profits and margins clearly spooked the market, with shares falling 24% before recovering to end the day only 13% off. This reaction was no doubt heightened by the vague commentary on how exactly the benefits would show through.

Rightmove did not announce disruptive changes to the product offering, simply what could be described as intelligent product developments. Looking across to similar UK-listed businesses as a tech analyst, there is hardly anything radical about what Autotrader, Trainline, Moonpig et al are doing with AI. That’s good.

Concerns seem to centre on three negative scenarios

In the Rightmove analyst meeting, the concerns appeared to centre on three negative scenarios that are common to many ‘technology enabled’ businesses that bring buyers and sellers together. First there is the idea that ChatGPT and the like could be used by consumers to create detailed intelligent searches that make Rightmove and its ilk redundant. Then there is the alternative tech-positive proposition that AI could be used to build and power a host of new competitors. Contrasting with these is the view that all this AI spend will add nothing in terms of revenues, profits or returns.

DIY – AI enquiries

The functionality and value offered by an AI enquiry and search can be substantial and go way beyond a simple Google search. Consumers could replicate much of the value created by marketplace companies for themselves with well-constructed enquiries and follow-ups, if they understood how. But there is a considerable difference between consumer ‘could’ and consumer ‘will’, as few things are more difficult to overestimate than consumers’ laziness.

A key factor in the value of the AI result will be the freshness of the data, and with general models only being updated every six months or so the gap is filled by internet search results. The resulting response and added insight is unlikely to match the incumbents’ services, particularly where they possess proprietary data and insight.

The DIY AI argument ignores the fact that these search / AI engines make their money on delivering users to places where they can make purchases – places like Rightmove, although in Rightmove’s case the vast majority of users come direct.  The AI general search business model could change this, but at present there is no clear divergence from its less intelligent forbear.

AI replication

The alternate tech-based bear case is that AI allows these businesses to be more readily replicated, creating a competitor or competitors.

With a good knowledge of how the APIs, regulations, markets and customers work, a suitable product could perhaps be created on top of a general model using augmentation of data to get towards something that an initial consumer user might struggle to differentiate from the original. However, in many cases the incumbents have, like Rightmove, rights over content and such information as customer behaviour and price movements.

Importantly, none of this AI effort creates a brand or mindshare; this is crucial in markets where most users go direct to the incumbents’ websites or, increasingly, use their apps.

Does it work and deliver sustainable returns

Not all these AI-powered increments to the product will deliver revenue, profit and value benefits, but that is normal – and that is why Rightmove’s comments lacked clarity.

However, from the history of these businesses it is evident that they can increase revenues and profits with new applications, AI-powered or otherwise. Of course, there is only so much of the value chains that these businesses can obtain, but in the case of Rightmove there would appear to be further opportunities to extend and grow in its current space, as well as into connected areas.

However, there is nothing to suggest that many marketplace business models or value chain dynamics are going to change substantially just because of a bit of AI, and in competitive markets we should expect to see some or even all of the AI gains competed away – as with any copyable innovation.  The increment announced by Rightmove was only £18m of investment/spend for a high-margin business with annual revenues of around £400m, so this really is only a bit of AI.

AI’s soft gains could disappoint before hard gains show through

This brings us to the big-picture issue of whether any gains from the current massive global AI investment will be sustained, or to be more precise whether the innovations, products and services enabled by this investment will deliver appropriate returns.

The primary AI effort for businesses at present is within sales and marketing, and the businesses most likely to be doing so are within technology, media, telecom and healthcare. Much of this effort is being undertaken in silos, defined areas of businesses, where the expected gain is greatest, easiest or readily demonstrable.

This suggests that in the short term there should be plenty of positive AI investment returns stories. It is not too radical a conclusion to draw that much of these soft gains will be competed away as competitors replicate the new AI products and applications.

It will be interesting to see whether the soft gains in the short term are competed away, thereby frustrating and disappointing investors, before we see AI generate more sustained hard gains from agentic AI and from such areas as pharma development, high-end engineering, plant management or public sector administration, where AI spend is more long term in nature.

Ian Robertson
irobertson@progressive-research.com
07768 276595