How to forecast your Stripe sales or revenue

Although various AI-based models for recommending content or making predictions have been known for a while, it is still tricky to figure which metrics are best to run through them in the first place. That of course, depends on the type of product you have and the business goal you have for the longer term.

If your issue is for example high shopping cart abandonment, you should dig into that problem first. You may then decide that suggesting more relevant items to you users will potentially increase shopping cart checkout. This can be done by implement a fully automated recommender model. This is one of the many advanced techniques that can squeeze out more from your conversion funnel.

But one of the most useful insights that deliver a lot of value is a “simple” forecast. For any business, running forecasts regularly should be common practice. One of the metrics you likely want to keep an eye on with a forecast is revenue or sales.

With Stormly you can easily forecast and find trends in more complex metrics such as a conversion funnel or user retention. It’s important to keep an eye on the trend of these relative measures as well, because they can point to underlying problems with your app, website, or conversion funnel.

In this guide, we’ll take a look at forecasting and finding trends in what is probably the most important metric for many online and offline businesses- revenue.

If you are using Stripe, like the rest of us, you are collecting quite some data already. Why not make use of it?

Why forecasting sales is important

You see a drop in sales for the last few days, and start to panic. By running a forecast you will know if this decrease in sales is expected or a real problem. A forecast may indicates that because of seasonality, your sales sales are always down 30% starting this time of the year. Even better, by running a forecast a few weeks ahead of time, you will know that sales is expected to be down 30% next month. When the time comes, you won’t be surprised and can for example bump up sales by acquiring more traffic during next month.

You can anticipate even more fine grained seasonal patterns- not just the ones that occur for a given year, but also more detailed ones, based on the day of the month, week, or time of day.

Planning capacity. If your sales is about to go up 20% next month, you can predict roughly when you’ll need to bump up your infrastructure. This will let you save on your operational cost and maybe, on the staff support cost as well.

Predicting the coming days or weeks and planning the advertising budget. If you expect the traffic to go down over a specific time period, you can run additional campaigns to help increase it and not miss out on sales.

With Stormly we’ve made it super easy for you to forecasting revenue or anything else, with just a few clicks really.

Step 1 - Importing your (Stripe) sales data

First you need to import your historical data to the platform. Stormy supports Excel and CSV files directly, which can be exported from your Stripe dashboard.

1. The first step is to sign up to Stormly: Click here to signup

2. Then log in to your Stripe account.

3. To export your data from Stripe, open the Payments item in the left menu bar and then click the Export button at the top right, as shown below using the links marked red:
   

4. In the export screen select “UTC” under Time zone and “All” under Date Range. Finally select “All” under the Columns:
   

5. Create a new Stormly project by clicking the Stormly logo at the top left. Then click “Manage Projects” from the dropdown menu. In the projects page click “+ New Projects” and then enter a project name as shown below:
   

6. After you create the project, you end up in the Setup Data page. First click “Import File (Excel, CSV, JSON)” here. Then as shown in the image below, choose your CSV file. Once you’ve done that you can select the columns P, A, D and N for the fields corresponding with user id, event id, timestamp and event name. Finally choose “Import Data” to start importing the CSV file. Wait a minute or two for your import to complete.
   

Step 2 - Getting your sales forecast

To open the Forecasting Insight, click the Insights top menu item, and then and open it.

The default Insight that opens is to forecast specific event actions, such as signups, payment completed, etc. We don’t want to do that now, so close the close cross at the top right of the questions modal.

Then on the left side open the Forecast Total Value. In the question modal that pops up, you can set the number of days to forecast and more. More importantly, select the dropdown 1st event property, then pick the Paid event, and its property Amount.

Now click the Continue button at the bottom. After a few seconds you will get your forecast, that looks something like below.

In the first section you see the actual forecast for the next 30 days, which you can use to anticipate your sales. On the right side a general trend without flucations is shown; in this case sales is clearly expected to go down over the next 30 days, as indicates by the green line.

In the sections under that, you can see that thursdays, saturdays and fridays have relatively more sales revenue than other days. In tems of patterns for an average month, it’s shown that the 11th day till the 20th day of the month are usually better in terms of sales revenue.

Next steps

You can become even more data driven by getting daily updates on your forecast, automatically. You can do so by clicking the thunderbolt icon at the top right, and setting the update interval.

But before that, you need to get realtime data ingestion setup with your project. To get that up and running, head back to Setup Data page via the Projects page.

You may need to ask one of the developer on your team to get you setup. It’s a trivial task that usually doesn’t take more than half an hour if you’re already using Stripe.