Profit per Pixel-Second? [Editorial comment from March 2009’s edition of Internet Retailing Magazine]

This article appeared in November 2008’s edition of Internet Retailing Magazine.

This last month has seen Ian Jindal up to his nose in web analytics and trading reports, pondering the painful question of “conversion”: can this really be a useful metric for etailers?

Now there’s no arguing that conversion is a ‘metric’. Number of visitor sessions ending in a purchase, divided by the total number of visitor sessions. Ta-daah. After a month of looking at ‘the math’ (as our US friends would say) I’m losing faith that this is an actionable or useful metric.

Conversion has its place in reporting: at a gross level it’s a rolled up indicator of “persuasive attraction”: how many people are being attracted to the site, and from that level how effectively are they persuaded to buy. However, if conversion is up or down then it’s not clear that there’s a single lever to apply. Unlike a car’s speedometer, it’s not straightforward to press or ease the accelerator to regulate the speed. Conversion as a metric lacks direct impact on the business: it’s observational rather than action-oriented or definitively comparable.

Two retailers may have the same conversion, but revenues many millions of pounds apart. Equally, a multicategory retailer may have very different conversion rates across products from £3000 sofas to £20 trousers. Who’s making the most money? The best use of resources?

Where products have a long consideration cycle it’s misleading to simply consider the final, purchasing visit as ‘effective’, while the previous ones are somehow ‘overhead’.

A final objection is that in a multichannel world research and purchasing my take place over different channels – where then does ‘conversion’ help drive our activity?

My contention now is that conversion should be consigned to the dustbin of pointless, but detailed, metrics – nestled alongside the “hits” measurement from the 90s.

What then can satisfy us as being a rolled-up metric, against which you can manage your business and which is directly comparable across etailers, categories and time?

Such a metric would need to include as a factor the notion of “profit” – otherwise it’s simply an engineering metric of ‘activity’. Equally, it should relate to maximising invariates.

In traditional retail we have a universal metric of profit per metre of shop space. This illustrates a retailer’s effective use of the fixed resource (space) and their management of yield and profitability.

Online, our limitations are different. While some of us may still believe in the misleading fallacy of the ‘infinite warehouse’, we know that our two limiting factor are:

  • the screen size – there’s no use having a magic warehouse if the window is small, dirty and germ-ridden!
  • the attention span of the customer, measured in the number of seconds they’ll spare for you in which you may persuade them to buy.

This leads to a metric that could be expressed as yield or profit per pixel-second.

We know that not all pixels are created equally and placement is vital: persuasive messages, imagery, promotional prominence, branding, tools – all vie to colonise the limited space. Different retailers take widely diverse options – luxury brand stories versus pile ’em high money-off screaming. Let’s measure their effectiveness on yield.

So then to customers. We’ve noted before that a rising tide floats all boats, and until the end of 2008 there were plenty of new customers spending evermore time each online to allow every retailer to record growth. However, in a saturated market there’s evidence that online customers are settling into a core group of a dozen retail sites (where ‘retail’ include aggregation/affiliate, voucher and cash-back portals who – from a customer’s perspective – are simply alternative ways to shop). The battle now is for the customer’s attention as much as for their money once you have that attention.

Of course, the cunning reader will realise that I’ve not included here the important offline dimensions – time spent handling goods in store, discussing configuration of complex furniture offers, speaking with experts about high cost items… However, a number of retailers are experimenting with ways to track a customer’s activity across channels – a voucher code in-store redeemed online, purchasing cards, logging, custom item codes, case numbers…

As these gain traction – and retailers can take a multichannel view of the effort and investment needed to support sales – then yield-per-pixel-second will morph into a metric of ‘yield per customer engagement second’, across all channels.

At this point we’ll have a universal, comparable, profit-oriented metric. This will allow us to benchmark ecommerce operations, but also see the value of etail within the mix – and importantly draw in the costs of contact centres, store activities and direct mailing into an overall cost of doing business. Maybe not next month, but such an index of effectiveness must be the aim.