Measuring the success of your app integration

This article is part of a series on how to scale and optimize you app integration on the QuickBooks platform. See Part 1, “Your app integration is a living product – measure it.”

In this article, we go through useful metrics to measure the health of a platform integration. Your preferred metrics may vary depending on how you interact with the QuickBooks platform. Chances are that your aim is to grow and retain your customer base. If so, best is to follow a funnel model. The aim is to optimize how to convert and retain as many of the 4.2M QuickBooks customers shopping for your app on the marketplace:

1. Discovery: “Are you getting enough eyeballs to your app?”

QuickBooks app card views: When you publish an app on the QuickBooks marketplace, customers can find you through your app card on apps.com or the apps tab in QuickBooks. This is the front door to access your integration and sign-up for your app. Use a Google analytics tracking ID in you app listing settings to track visitors in real time.
Landing page views: if you create your own landing page so customers can find out about the integration to QuickBooks, check its effectiveness in attracting visitors. Some customers don’t go to your app or the app store to look for integrations, but rather Google it: “Integrate X with Y.” Consider using Google AdWords to further increase traffic. See our Search engine marketing and advertising guidelines for QuickBooks apps.

While not being true discovery metrics, the following two data points are also important to monitor as they greatly influence the discovery of your app on the marketplace:

QuickBooks star reviews: The marketplace uses star reviews as a one of the many criteria to rank your app in search results. An app with 1 star review will typically rank lower than a similar app 5 star review, thus obtaining less share of traffic. Monitor how your star rating is evolving over time.
QuickBooks review ratio: Number of reviews / total active users of the integration. In combination with the feedback rating, the number of reviews also influences your search ranking. You should aim to collect feedback for at least 5% of active users. Remind your customers to leave a review about your app using your review link: https://appcenter.intuit.com/app/apps/appdetails?tab=review&shortName=<yourapp>.

2. Onboarding: “How many customers enable the integration?”

Leverage the marketplace as an acquisition tool by redirecting new customers to a streamlined sign-up flow. Not every customer who comes from the marketplace is a new user. Some are already customers of both platforms, looking to activate the connection between the two systems. In this case, your app should direct them to a sign-in flow to link QuickBooks to their existing account.

New vs. existing customer ratio: Compare the number of users connecting your app and going through a sign-in vs. sign-up flow.
Business vs. accountant ratio: QuickBooks is a two-sided platform allowing businesses and accountants to collaborate and connect apps. If you have a significant proportion of accountants connecting your app to their clients, you should consider designing a dedicated and tailored onboarding flow.
App listing to landing page traffic conversion: Number of unique visitors to your app card / number of unique visitors to your landing page. A low conversion is a sign your app card does not effectively sell the benefit of your app. Put your marketing hat on and check if you can improve the content on your app listing. Check these tips to fill out your app card.
Sign-up conversion: How many new customers coming from QuickBooks complete your signup flow. If conversion is low, there could be an issue with your signup forms. Try to reduce the amount of fields required to complete sign-up. Pre-fill the info using the Intuit QuickBooks user profile data available through OpenId connect to make your signup a breeze.
Setup conversion: Once a user is signed-in or signed-up, do they complete the steps to activate the integration? Depending on the complexity of your product, you may have multiple steps as part of the setup process. A conversion below 70 percent is a sign customers are getting lost. Consider defaulting some of the settings. Avoid technical or accounting jargon such as FIFO, COGS, Ledgers, or Journal entries.

3. Activation: “How good is their first experience with the integration?”

Even after enabling the integration, the customer might not be using it and perceiving its value, whether that’s data being synced, saving time, or money. For example, if your app extracts accounting data from QuickBooks to generate sales insights, the a-ha moment is seeing their sales performance dashboard for the first time. Identify key actions powered by your integration, then track whether customers are completing them.

For apps that sync financial data in and out of QuickBooks, the value is getting all the data successfully updated in QuickBooks. Check the success rate posting this data in QuickBooks – see #6 below, API integrations metrics, for more info.

4. Retention: “Do they keep using the integration after a few months?”

In the context of an app integration, retention means the number of users who still have an active data connection between Quickbooks and your app, and continue to perform key actions over time.

Usage retention: Monitor the key actions users can perform and measure the percentage of users who complete the event over 30, 60, and 180 days.
Active connections: Number of active connections to QuickBooks over time. This metric is available through your app dashboard. When a customer grants your app access to their QuickBooks data, QuickBooks generates an OAuth token. This token represents the active connection between QuickBooks and your app. A token can expire for the following reasons:
- Customer manually disconnects your app from QuickBooks via the my apps menu.
- Customer manually disconnects QuickBooks from your app via a disconnect button in your app.
- The OAuth refresh token expires after 100 days. This happen if your app is no longer submitting API requests to QuickBooks due to user inactivity, and you have not implemented a token maintenance schedule to obtain a new refresh token. See token expiration.
- Customer either cancelled his QuickBooks subscription, or did not upgrade from a trial to a paid plan after connecting your app during the trial period.
- Customer cancels his subscription to your app.

5. Output metrics: “Do we generate user or revenue growth from this app?”

Net user growth: Number of new users acquired from QuickBooks minus churned customers over a given timeframe.
Direct revenue: This can be direct revenue; for example, if you charge a setup fee, a fee for the use of the integration, or a referral revenue if you have a referral agreement in place.
Indirect revenue: Even if you don’t directly monetize the integration, there is an indirect financial impact for SaaS subscriptions which can be quantified using the Lifetime value (LTV) uplift. LTV is the average revenue a single customer is predicted to generate over the duration of their subscription. Calculate the LTV for customers not using the integration and compare this with the LTV value for active users of the integration. The simplified calculation is LTV ($) = average monthly revenue per user / monthly churn. If the integration delivers customer value, you should see a positive difference. In other words, the integration makes your customers “stickier” and more likely to stay subscribed to your app.

6. API integration metrics: “Is it working?”

At the core, an integration depends on the quality of the data being exchanged by the underpinning partner API. Users expect their data to flow freely, in real time and without fail. This is especially true when dealing with financial information, which businesses rely on to take important decisions. Proactively tracking the following metrics will help you identify issues before your customers do:

API error rate: Number of API calls not returning a http status 200 / total API calls. A ratio above 3 percent for a given timeframe indicates an issue with the integration, or a server downtime. Investigate further.
Percentage of users impacted by an API error: If an error is isolated to one customer, it could be due to a specific setting or data input which may not warrant urgent attention. If it impacts a significant portion of your users, investigate and fix urgently.
Top 10 API error codes: Create a monthly report summarizing the errors returned from the QuickBooks API responses. To help with troubleshooting we compiled a list of common error messages and how to handle them.

To track API level metrics, you will need to generate application logs whenever an error is returned from the QuickBooks API. There are a few tools out there which provide a visual interface to monitor and track errors such as ApiGee or Splunk.

7. Pair your data with qualitative feedback

The data collected only gives you one side of the story. Metrics in the red are symptoms of an issue with your integration. You need to talk with your customers and understand how they use the integration to pinpoint the cause. Are they finding it easy to set up? Are they getting all the data they need to complete the task in the other system?

At Intuit, we often visit our customers in what we call “follow me homes.” We meet them in person at their workplace, and observe how they use our products alongside other apps. Often, we see app integrations which miss crucial details. For example, an app which only syncs summarized transaction data, causing incorrect sales tax calculations if some line items are not taxable.

There are many other ways you can collect additional qualitative feedback:

Prompt your users to leave feedback on your integration past a 60 days period, then do the same a year later to monitor any changes in sentiment.
Display a feedback form in your UI to collect the reason if a user disables the integration.

8. Tracking tools

Here are some of the tools which our app partners use to track the health of their QuickBooks integration:

Hotjar: Heatmaps – let’s you see scroll mechanics i.e. how far people go down the screen and heatmaps.
Google StackDriver: performance and diagnostics data (monitoring, logging, tracing, error reporting, and alerting).