Usage Trends

clustering customers and comparing the performance inside each cluster can help you better understand their health however, if your customers are different in terms of size, industry, subscription type, or maturity, clustering may become irrelevant if you find yourself in such a scenario, we recommend that you compare each customer’s performance individually and set alerts in case the activity drops below certain levels to achieve this, build a trend that compares relevant timeframes (e g month over month, week over week) and tells you if the usage trend is positive or negative to build the trend, you can define https //kb custify com/cuwi calculated metrics and use them in https //kb custify com/health scores to observe the evolution of an event occurrence over time here’s how step 1 build a calculated metric using the formula trend = (time period now / time period before) 100 – 100 here is an example of a trend that compares the number of logins from march with the ones from february no of logins february no of logins march trend health score 100 40 60% red 100 80 20% yellow 100 150 50% green in the first row, you can notice a decrease of 60%, meaning that the customers logged in less than they did before in the second row, you can notice a smaller decrease of just 20%, while in the third row, there is an increase of 50% this is how you build the formula in custify ((event over time(‘login’,7) / (event over time(‘login’,14) – event over time(‘login’,7))) 100) – 100 step 2 create a https //kb custify com/health scores using this calculated metric here, you can add the metric you just calculated and define the score intervals for example, 80% can be the worst value, and +20% can be the best value red (bad) — a decrease larger than 50% yellow (average) — between 50% and 20% green (good) — an increase or a decrease smaller than 20%