The Economics of Selling on the Internet

Internet retailing has been a fascinating long-term case study in technology disruption where, two decades in, initial bullish expectations regarding the end-state unit economics of the industry have yet to come to fruition.

The fact is that even today in 2023 it remains extremely challenging to find a consistently profitable, large-scale internet retailer that generates a satisfactory economic returns on capital employed.

So if the economic benefits of internet retailers “bypassing the middlemen” and “avoiding fixed rent expenses” are not observable within the industry profit pool, then where are they being realized? Is any economic surplus being captured by anyone (apart from consumers)?

Where has disruption played out?

Firstly, let’s look at where “disruption” has and has not played out within the broader retail industry value chain.

People have long had the ability to order something from their home and get it delivered to their home – whether through mail, telephone or a door-to-door salesman. Whilst the internet has made this faster and more efficient, it has not been game changing. The logistics of getting a physical product from a retailer to its customer remain labour intensive and basically subject to the laws of physics (see Milkrun).

In fact, most of the time, getting your customers to collect their purchases from a distribution point nearest to them (in other words, a bricks-and-mortar retail store) is more efficient for the retailer than to have to deliver each item individually to each customer (Interestingly, Costco seems to have taken this to the extreme and made it attractive for customers to drive reasonable distances to collect their purchases from Costco’s warehouses).

The real game changing “disruption” to the retail industry has been in customer demand generation.

Beginning with simple display advertising (e.g. banner ads, pop-ups etc in the 2000s), internet advertising has evolved over time into hyper-efficient, real-time, auction-based, pay-by-performance platforms dominated by Google and Meta.

These platforms effective offer “customer demand as a service“.

And today “customer demand as a service” is highly democratized. Anyone with a personal credit card and 10 minutes of time can compete directly with the largest retailers in the world for incremental customer demand through a self-serviced platforms.

Over time, each and every online advertising niche have become hyper-efficient. In fact, players such as Tradedesk employ “traders” to take advantage of market arbitrage opportunities just like traders would in the financial markets. In such scenario, it is inevitable that any economic surplus eventually gets competed away.

As the technology got better, the outcome for users got worse

So we effectively ended up in a counter-intuitive scenario where the economic outcome for users of internet advertising got worse as the enabling technology got better.

The auction-based competition facilitated by these platforms ensures that an advertiser’s marginal cost of customer acquisition closely tracks its gross profitability and in turn makes it very difficult for operating leverage to be realized even with scale.

Another way to think about this is that if your business is dependent on internet advertising to generate customer demand, your competition for a share of the economic surplus is primarily against Google and Meta, rather than against your direct industry rivals.

The only way for an internet retailer to get off this hamster wheel is to be able to either generate their own customer demand independently (not easy – you basically have to get to Amazon-scale) or work out how to materially expand your customer LTV (e.g. increasing repeat purchases, expanding ARPU etc) so that you can afford to pay more for traffic versus competitors.

The Australian listed internet retailing landscape

Looking at the universe of listed internet retailers in Australia, it is evident that the industry have had to pay materially more to generate each dollar of revenue today compared to 5 years ago (even COVID-19 where the industry enjoyed unprecedented organic demand did not break this trend). Their operating leverage actually decreased rather than increased with scale.

Redbubble’s (ASX:RBL) paid advertising expenses as % of revenue has consistently tracked upwards at the expense of its own profit margin. This would indicate that despite its touted “network effects” and “long tail of niche creator content”, Redbubble’s economics at this point in time basically tracks closer to that of a retailer of mugs and t-shirts.

Kogan (ASX:KGN) has also seen its percentage of revenue spent on direct marketing increase rather than decrease over the past 5 years. Whilst there are some promising signs of a reversal during the latest half, it is caveated by the discounting that took place during the half (where Kogan saw an accompanied a ~290bps decline in gross profit margin).

Adore Beauty (ASX:ABY) literally stated that it had “lower operating leverage” in its 1H23 investor presentation.

Temple and Webster (ASX:TPW) seems to have fared better where its dependence on “customer demand as a service” declined over the most recent half year and it was able to achieve headline profitability (I wrote about Temple and Webster right after its IPO in 2016).

So why is everyone paying higher customer acquisition costs to generate each dollar of revenue? And who does this spend accrue to? Well it’s interesting to observe that over the same period, Google’s advertising revenue has more than doubled (off an already very large base).

An inefficient market is your friend

Interestingly, whilst it hasn’t exactly been a smooth ride for bricks-and-mortar retailers either, it seems that many more of them are able to generate a consistent profit at scale compared to their internet counterparts today.

The key here is that pricing for demand generation in bricks-and-mortar retail – i.e. foot traffic paid for in the form of retail rent – is relatively static. The suppliers of your customer demand (i.e. your landlord) do not demand vastly different rent per sqm for different uses of the premise. And once you sign your lease, the rental costs are relatively fixed over multiple years.

An interesting nuance here is that it is structurally difficult for retail property owners to shift towards a revenue model that more closely tracks a retailer’s revenue (i.e. turnover rent) given their own investor base. Property debt financiers and investors typically value a stable rental yield over and above less stable equity-like returns (if they wanted such exposure they’d just invest in the shares of the retailers).

So if you’re a retailer that can monetize a given retail property better than others (e.g. you may be the category leader with national scale like JB Hifi), you get to keep the upside from this “outperformance” and in turn generate economic value-add for your shareholders. This is very different to what is experienced by internet retailers, where Google and Meta puts up new competitors to bid against you in real time for any incremental traffic as soon as you find pockets of profitability.

Ultimately, being in possession of a network of localized moats is a much better place to be compared to being in a larger but almost perfectly competitive market with low barriers to entry. It’s the proverbial choice between being the big fish in many many small ponds versus being a small fish in a very big pond.

Technology can be a double edge sword

The technology that enabled the internet retail industry became so good at what it does that it created hyper-efficient markets and brought about fierce competition to the detriment of its own users.

In other words, the business models of internet retailers have themselves been disrupted by the very technology that they were supposed to utilize to disrupt the bricks-and-mortar retailers.

I previously wrote about bricks-and-mortar retailing here.

Note: The above article constitutes the author’s personal views only and is for entertainment purposes only. It is not to be construed as financial advice in any shape or form. Please do your own research and seek your own advice from a qualified financial advisor. Being obviously passionate about investing, the author may from time to time hold positions in the aforementioned stocks consistent with the views and opinions expressed in this article. Disclosure – I hold no position in any of the stocks mentioned in the article at the time of publishing this article (This is a disclosure and NOT A RECOMMENDATION. Any position held by the author may, for example, be part of a pair trade or a hedging position so it is not informative in isolation).