Developing Startups Part 3: Measure

This post is part 3 of a 4 part series “Developing Startups: Lessons Learned Over Two Decades”. If you’d like to read it from the start, you can jump to the first post here.

Measure how customers respond to your product/service. This is where analytics come in handy.

Analytical Tools

Analytics are a requirement for any startup. They help you analyse user behavior, allowing you to make informed decisions in regards to improvements to application flow and functionality. They are also an invaluable marketing tool, and necessary for reporting growth to investors and stakeholders.

When you’re first starting Google Analytics is the way to go. Free and packed with features, it’s a staple not only for startups but pretty much for all sites. Basic page tracking can be used straight out of the box but should be configured for event tracking and and ecommerce tracking if your startup has a cart.

For mobile apps you should use GA and at least Crashlytics for crash reporting (also free). With so many different mobile OS and so many different sets of hardware, it’s impossible to be able to catch or even recreate many bugs that your user will be experiencing. This is were Crashlytics comes in. It’s a life saver.

It’s always important, through the life of a platform, to have good analytics. But as a startup you should take advantage of all the free tools that exist and let your analytics grow with you. Here’s a diagram that shows a stage by stage overview of an analytics growth plan.

Within your analytics there are many things you should be looking at. Here are a few:

• Funnel Analysis - Whether you’re selling products or trying to increase membership there is usually a goal or set of goals that you want to achieve in regards to user interaction. With these goals in mind you can set up funnels in Google Analytics to help visualize your user flow. This helps find drop off points along the funnel that can tell you where users exit desired path. Now that you can pinpoint problem areas you can start start planning solutions.
• Cohort Analysis - A cohort is a segment (group) of users based on a date. For example, if you’re looking at acquisition date cohorts by user retention, your analysing the user retention of users acquired on a certain date. Acquisition date cohorts by revenue per user means you’re looking at how much revenue was created over a period of time by a group of users acquired on a certain date. Here’s a good post that helps you understand Google Analytics cohort report.

Split (or A/B) Testing

This is a great tool for gathering information on user preferred functionality. Forgive me for making this point again, but you’ll be surprised by how often you’re gut can be off in regards to expected user functionality. A/B Testing lets you settle the question quantitatively.

Server/Error Log Data Analysis

This has less to do with measuring user response and more to do with measuring infrastructure response to users. Both when a system is launched and later when it’s being maintained and improved upon, the infrastructure has to be carefully monitored in order to be able to preemptively address issues where possible.