Modelling eCommerce revenue (Forecasting eCommerce performance)

We’re working on quite a few eCommerce projects at the moment.

A handful of the websites are for large retailers that have never invested in eCommerce, and so really have no idea of what to expect in terms of eCommerce revenue. Offline, they know their margins, how many customers walk through the door and the average transaction, but does this guide us when forecasting for web?

Not really unfortunately; knowing the margins and cost centre around a product help to estimate the profitability of the product itself – once sold – though even that is tricky, because whilst we eliminate offline costs centers such as rent and sales staff, we introduce complexities and costs that don’t directly translate offline: the cost of sale formula is changed.

When we help clients with forecasting their eCommerce revenue and profitability, there is a handful of easy parts to the model, at least in terms of knowing they exist:

• Number of users (often a guess based on a static website)
• Number of these users that will start adding product to a cart (%)
• Number of users that will ultimately purchase (% of total carts)

The unknowns are usually what it costs to drive the customer to the store in the first place, and the value of the cart; though this said, my experience is that the average cart size is always impressively larger than expected.

Point of the story is that we’re just about to complete what I believe is the best model in Australia for modeling eCommerce revenue based on all the various inputs and outputs. It can work forwards or backwards; you want $2m in sales, here is how many customers you need and the cost of getting them; you have 50k visits, here is how much you should expect. We are going to adapt it to different industries as they come to us.

It is exciting to be able to provide such surety to clients in so far that at least they understand there are economics underpinning it all. Offline dynamics slowing leak online.