Forecasting for a bullish future
The customer:
The customer is a leading provider of password management for both consumers and businesses - they represent a popularization of cyber security that has ballooned in importance and value recently, a market valued by Mckinsey at $2 trillion in addressable market. Our customer provides a user friendly and intuitive solution to password management, allowing for secure, single-use passwords across all accounts a customer has online; this ensures that no single account breach can expose a customer to widespread vulnerability among their remaining online accounts.
The challenge:
Our customer was scaling it’s marketing capabilities and budgets, with increased exposure in paid search ads, display networks, and affiliates - this was made possible with a recent (and substantial!) round of private investment funding. In order to manage expectations and push internal teams for higher performance, the customer needed to scale it’s ability to forecast future customer acquisition numbers. These forecasts needed to serve multiple purposes:
🎯 as targets for internal marketing teams
📈 as inputs to financial revenue models
⏲ as a benchmark for diagnosing issues in the marketing funnel
Goals as a diagnostic tool
Goals aren’t just aspirational - they tell you when a part of your plan is falling short, so you can fix it or fix your expectations
Solving the challenge:
This challenge is a result of scale outpacing instrumentation. To solve this problem, we planned intelligently for a fuzzy future, so that the customer could have an early warning signal if expectations began to fall short.
Expectations and assumptions are key when establishing forecasting.
Stakeholder interviews revealed that legacy forecasts were not usable for diagnosing slowdowns in the marketing funnel. The cause was a lack of granularity in the forecasts, combined with inadequate assumption documentation - the customer may be missing or exceeding it’s legacy forecast - but leadership could not articulate how or why
This required a scalable framework for forecasting, where increased granularity could be included, allowing for both quick diagnostic of performance issues and roll-ups to top-line business targets
Stakeholders required increased visibility into location of signup (app or website), marketing channels, and monthly date granularity. We made a recommendation to leverage daily granularity for forecasts, allowing for precise diagnosis when an area of the business begins to lag behind goal
We sourced historical data for our customer’s business, in order to forecast forward for 1 year
Daily website session volumes were leveraged with open-source Python forecasting tools to extrapolate expected future session trends - seasonality, trend and margins of error were forecasted. Similar efforts were conducted for conversion rates at key stages of the customers' funnel
Finally, we created a self-serve dashboard for our customer, summarizing actual performance versus forecasted values. This dashboard supported both summary performance readouts, as well as granular self-serve information, should the customer want to dive into under / over-performance drivers
The results:
The output of our efforts was a forecast that accurately summarized business performance over the next quarter, with a margin of error less than 5%. Our forecast had a 3x improvement in diagnostic capability versus legacy forecasts - this was due to the inclusion of forecasting by additional dimensions at the channel level. Finally, our forecasts supported daily granularity, versus legacy monthly models. This resulted in up to 3x improvement in detection times for funnel issues.
Ready for the next step?
Reach out and start a conversation - we won’t pressure you, give a hard sell, or send endless follow-up emails.
