![]() ![]() We admittedly focused less on these fixed cost elements, which are harder to model in the abstract. To model Operating Expenses, we included assumptions for monthly headcount and salary, as well as other line items for overhead. ![]() The product of these two rows creates the Gross Product Revenue line item at the very top of the P&L.įrom here, we added various assumptions about shipping revenue/costs, return/discount rates, and COGS components - each as a % of AOV. We built the Income Statement, or P&L, on the tab called “Monthly Model.” Under “Key Operating Metrics”, we flowed through Total Transactions and AOV from the “Cohort Model” tab. We also placed the average order value (AOV) assumption on this tab, which will play into the calculation of revenue on the P&L tab. 3-year LTV and LTV/CAC are also calculated here, but require input from the P&L. You can see some other calculated values in this tab as well: weighted average CAC (aka “blended CAC”), 3-year cumulative transactions, and repeat rate. Moving from “cohort time” to “calendar time” by summing along the hypotenuse of the Jun 2018 cohort. You can see how the cohorts stack up in the classic “triangle view” below, with month 0 justified left: We also allow for some improvements in that baseline retention for subsequent cohorts. “In month” retention is equivalent to the percentage of the cohort which shows up to purchase in a given month. Note the model uses “in month” retention and not cumulative retention. After the month of first purchase (=month 0), they appear again along a baseline retention curve that levels off to a steady state. These new users then follow a standardized cohort behavior. Adding paid to organic users yields the total new users in the cohort. We account for this “organic halo” by assuming a ratio of organic to paid users for each cohort. In a healthy business with good word-of-mouth, new paid users tend to drag a few organic users along with them. ![]() Together, they yield the number of new paid customers per period. We start on the “Cohort Model” tab with assumptions on monthly customer acquisition spend and customer acquisition cost (CAC) through paid channels. We summarized our approach below and left a more detailed user guide in an appendix for the extra curious. We haven’t seen anyone provide a solid framework for integrating cohorts into a forecast so, my partners Jeremy Liew, Natalie Luu, and I built one with fictitious data. We’ve previously written about how we use cohorts to estimate lifetime value (LTV), and how graphs of cohorts (aka “spaghetti graphs”) are the best tool to understand long-term customer engagement. Cohorts are ground truth for e-commerce.Ĭohorts are the atomic unit of analysis and forecasting at Lightspeed. We think there’s a better way to forecast it all starts with cohorts. Some forecasts rely on arbitrary assumptions, like monthly revenue growth. Part of that struggle is the inherent uncertainty in any early stage business, but another major factor is the structure of the model used. Nearly all of them struggled to forecast financials at the earliest stages. Blockchain gaming and the metaverse have also emerged as new trends in the gaming and crypto sectors.A few of Lightspeed’s e-commerce portfolio companies.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |