observable logo

Where to Start with Data Analytics for your SaaS Business

The key to creating an effective strategy is to prioritize implementing each business initiative and metric collection mechanism by identifying the needs of your business, and the phase of growth you're in.

5 min read • Published

Lindsey Renken

Co-Founder and Chief Data Scientist

So you've heard you should be analyzing data for your startup, but you aren't sure where to start. This guild will help give you ideas and think through what's right for your business.

Businesses that are relatively new are typically still finding their sustainable marketing channels and becoming more aware of how to match customers with the key value proposition of their product. In other words, they are still finding product-market fit. 

At this stage, you’ll want to start by measuring how well you’re marketing to customers, and also how well customers are responding to the initial product.

Afterwards, think about how to ensure customer success by defining KPIs for customers on-boarding and engaging with key features.

Lastly, make sure customers are having smooth and reliable experiences with your platform.

Data Collection

When prioritizing analysis, it's good to think about what data you'll need and when. Some data takes a lot of time to collect to be of much use, so you'll want to balance starting to collect data for future analysis with collecting data you'll need immediately.

  1. Collect survey responses from visitors and customers where appropriate. Measuring your marketing channel conversion rates by collecting the UTM web referral attributes for each customer session, and documenting all marketing copy topics covered will also be helpful.

  2. Next, you likely want to collect the metrics relevant to your on-boarding flow once you have one and maintain metrics for each step you create or modify. These include: number of steps, step success, length of time on each step, and any drop-off.

  3. Following, you’ll want to start measuring how users interact with the features of your product. Start by collecting the metrics on your primary value-adding features. Then, add metrics for your secondary value-adding features, then the rest of the product - including any steps users take to cancel their subscription (churn steps). Metrics to collect include: length of time spent using a given feature, number of times a feature was used on a given day.

  4. In addition, you will then want to start looking at bug collection tools your engineering team uses, so you can become aware of any issues customers are encountering. These tools will record the type of bug and related details. Try to link these to the exact user that was effected.

  5. Customer support tickets are good to collect. If you use a service, typically, they'll give you information on who handled the ticket, how long it was open for and who it was for.

  6. By default, if your team is using a service like Stripe, or Braintree, you’ll have payment event data collected on your behalf. You’ll of course want to use this valuable payment data for further analysis later on.

Analyzing the Data

After you have mechanisms for collecting your data, you can work on analyzing it.

  1. For early-stage companies or companies that have not done so already, your marketing team will want to first analyze which channels are bringing in the most customers, and which customers are getting the farthest in your funnel. Also, evaluate which topics are most effective, and determine the most impactful times to publish content.

  2. Your product, customer success and growth teams (or the founders if they are currently working these roles) will want to look at how customers are doing in the on-boarding flow. Analyze how time to completion of each step, and drop-off rates at strategic intervals change over time.

  3. Engagement with certain features - frequency of use and how recently they’ve been used by each customer - will become increasingly important further into the customer journey. Benchmark engagement for each customer against customer segments by looking at engagement across multiple customers.

  4. Product and engineering teams will want to know which features made the largest impact, so you can start to segment your customers based on what features they engage with most. Check if certain persona attributes are correlated to each group. These discoveries could help you uncover important trends and connections.

Set Actionable Alerts and Automated Triggers

Automation and alerts are good when you've systematized your processes. When you think you've make significant progress and certain actions are repeatable, start to think about automation.

  1. Marketing triggers like newsletter signups or engagement with certain pages or links on the marketing website that can be attributed to an individual are good to follow-up with via an automated message. The more timely the better with these interactions, in order to engage with potential customers when it's in-context.

  2. Actioning the on-boarding flow and improving feature engagement is important to prioritize because if you get these steps wrong, you’ll invariably have churned customers, and timing is everything.

  3. You can then start creating alerts for high engagement or disengagement, and measuring the level of churn risk customers have of based on how that customer engages compared to all of the customers in their segment. These alerts should trigger automated campaigns.

  4. You can also monitor the number of support tickets, bugs, and negative interactions with the product, marketing material, or the brand and report when customers are having a favorable or unfavorable experience.

These are some ideas to get you started.