September 06, 2006

E-Commerce Productivity

An executive at an e-commerce pureplay asked me a question about how best to measure the productivity of his website. He mentioned that his business experienced a significant increase in traffic, year-over-year, but experienced a reduction in conversion rate, year-over-year.

One way to analyze the sales productivity of a website is to take the Internet Retailer Top 500, remove outliers and smaller companies (leaving 251 observations) and plot demand-per-visitor (y-axis) against average monthly traffic (x-axis).

The scatterplot indicates that there is a diminishing rate of returns. As traffic increases, the productivity of each individual visitor decreases.

The relationship can be modeled using a "power" function (the word "demand" is used in this post, however, the relationships actually depict net sales, not demand, FYI).
  • Demand per Visitor = 816.2 * (Monthly Visitors ^ - 0.3754)
Using the relationship, we can predict an average Demand per Visitor, given average monthly traffic volumes:
  • 100,000 Monthly Visitors yields $13.33 per visitor.
  • 250,000 Monthly Visitors yields $9.61 per visitor.
  • 500,000 Monthly Visitors yields $7.50 per visitor.
  • 1,000,000 Monthly Visitors yields $5.85 per visitor.
  • 5,000,000 Monthly Visitors yields $3.29 per visitor.
  • 10,000,000 Monthly Visitors yields $2.57 per visitor.
Clearly, as a "brand" increases traffic to the site, it attracts traffic that will spend incrementally less (be less likely to convert to a purchase). Like all things in life, there is a law of diminishing returns. It is possible that my executive friend was observing a trend that is perfectly fine (traffic increase, productivity decrease).

The equation can be used to evaluate your e-commerce productivity. If you have 500,000 monthly visitors, you can expect $7.50 per visitor. If your site has $10.00 per visitor, you may be doing better than average. If your site has $5.00 per visitor, you may be doing worse than average.

Ah, the magic that occurs when you mine that data!!

No comments:

Post a Comment