Sources

  • E-Commerce Website Optimization - Dan Croxhen John, Johann Van Tonder

Profitability

Revenue = AOV* x NoT**

* Average Order Value ** Number of Transactions

  • CAC - Customer Acquisition Cost Average cost to acquire one new customer (‘cost’ as in marketing cost)
  • LTV - Life-Time Value, (a.k.a: CTV Customer Lifetime Value) How much a customer is worth to the business (in terms of revenue), over time

By finding the LTV, we can offset future CAC costs from that value, giving us an upper limit of how much CAC is justifiable w.r.t how much value they will likely provide the business over time.

Factoid: Many businesses do not even break even on their first transaction with a new customer after CAC costs are given!

Conversion

Think of a potential customer on your site as a sales conversation happening between a virtual salesperson and a physical customer

What do customers do in physical stores?

  • Try out products
  • Look over products in detail
  • Ask questions about: delivery, product durability, product effectiveness, ask for recommendations from the sales assistant etc.

Optimization vs. Experimentation

  • Optimisation: Weighing up multiple (new) alternative approaches (i.e: 2 new versions of a checkout page w.r.t performance)
  • Experimentation: Weighing up an alternative approach (i.e: new checkout page) against the original
    • A common methodology employed in experimental testing is the A/B test, referenced below

Marketing Testing

Marketing testing may look at metrics such as:

  • CTR

  • Conversion rate

  • Split Testing

  • A/B testing The original is employed 50% of the time (a.k.a: The ‘Champion’), and an alternative is employed the other 50% of the time (a.k.a: The ‘Challenger’)

  • Hold-Out testing A control group within the customer base is defined, and these people are excluded from any new campaign that is to be tested, to ensure that any results from alternatives can be evaluated as being better than before

  • Split Testing The same testing procedure as A/B, but using 2 alternative approaches (no control group)