What are the most effective ways to determine the lifetime value of a customer?

🍋 Imagine you have a lemonade stand, and you want to know how much money you'll make from one customer over time. → Average lemonade price: Let's say you sell lemonade for $2 each, on average. → How often they buy: Some customers come every day, some every week, and some once a month. On average, let's say they buy a glass of lemonade every week → How long they stay loyal: Think about how long you think a person will keep coming to your lemonade stand. Maybe a year? Now, we do some math: $2 (average price) × 1 (how often they buy in a week) × 52 (how long they stay loyal in weeks) = $104 So, one customer is like a cool $104/year! Now, you can decide how to make them even happier and keep coming back for more lemonade and fun! 🍋😄

Create Cohorts, otherwise you will always have a moving target. Average lifetime revenue per customer multiplied with the time you have them is your Customer lifetime value. It sounds easy but it can be deceptively hard to calculate depending on the complexity of your product, so make sure you simplify the calculation to a point where you can use it to make better decisions in your company.

Los análisis de cohortes son poco utilizados por la complejidad de realizarlos pero nos dan una gran información sobre todo en productos de software dónde el mercado se mueve de forma muy rápida. En productos con grandes costes de cambio quizás no es tan importante pero nos permitirá ver las tendencias y hacer buenas proyeccciones. En cambio, en productos con bajas barreras de cambio y mucho movimiento entre productos de softwares similares nos hará conocer de forma efectiva el LTV de nuestros clientes.

This will confuse things. Lifetime value should not be based on any forecasts. If you add prediction into the model then you need to make sure it is clearly stated. Predicted LTV can take in predictions as data points, but LTV needs to be based on historic numbers. Both numbers should be known and used, for different things.

Pedir feedback y hacer encuestas nos permitirá hacer buenas proyecciones. En una gran empresa de software en la que trabajaba éramos capaces de predecir el número de upsells y de bajas con gran acierto a partir de las encuestas de NPS, el número de llamadas a customer services así como la puntuación de la llamada en cuestión.