Making the Most of Your Analytics

Diana Smith of Segment

Diana Smith of Segment tells us how to get the most out of our analytics tools. In the pursuit of trying to be data-driven, we have been conditioned to track everything. Diana tells us why this can be dangerous if we want to draw useful insights from our data.

  • Here's what to listen for:
  • 00:49 What specifically are we talking about when we are talking about analytics in this context?
  • 01:55 What is the difference between user path and funnel tracking?
  • 03:10 Are there tools similar to Kissmetrics Path Report tool?
  • 03:57 If I’ve got my own database, why should I be using some sort of other analytics tool when I could just easily track events that happen on my database as it is?
  • 05:43 What events should I be tracking?
  • 07:22 When I set up what these events are, does it matter how I name them?
  • 08:19 What is the best naming convention?
  • 09:14 Why should I start only with just tracking a few events?
  • 10:53 What kind of info should I be putting in these properties?
  • 13:16 How do you connect and keep track of the who the referrer is? How does that work?
  • 14:45 How important are user demographics for data and tracking?
  • 19:59 How should I make use of the data that I’m collecting?
  • 23:25 Do you recommend that people create a bunch of accounts on these different sites and then choose one? How do you deal with the paradox of choice?
  • 24:57 What types of other analytics tools are out there?
  • 26:33 How do you decide which of these tools to use? What sort of questions should I be asking around “which one is right for me”?
  • 31:54 Are there any can-not-ignore metrics?
  • 33:03 How do I actually try and make a connection between the action that I’ve taken and the results that I’m seeing?
  • 34:34 Do you recommend waiting and focusing on qualitative things over quantitative?
  • 36:42 Is there a number you should be looking for in terms of when things should be statistically significant?
  • 37:19 In terms of doing the qualitative work that you talk about, and maybe trying to use quantitative data to make it match up with the qualitative data, or at least help … are there any specific strategies that you recommend for going out and getting that qualitative data?
  • 38:51 Let’s say that I am collecting enough data at this point. Even though I have a baseline for myself and my company, how do I know whether or not that’s a good baseline?

"It’s more about awareness. It’s about telling you okay, what’s currently going on with my customers."

"I’d question that it’s easy to track analytics to your own database."

"I like to break it down into three core parts of the user experience that are important to track."

"Only add more events as you have more questions that you can’t answer."

"It’s a lot of noise that won’t help until they have a question that it's is actually useful for."

"We are big sticklers for naming conventions, because we’ve seen what happens if you’re not strict."

"Things are changing today. Because you have this event data, it’s easy to make more targeted campaigns."

"This type of data that discovery/engagement/conversion information is super useful in a lot of tools."

"The other cool thing that you can do with this data is use it in e-mail tools."

"It goes down to like when you’re actually in the tool, which one do you find to be more useful."

"How do you focus on the tools that matter when there’s all of this choice?"

"Run it like you’re back in 5th grade and learning about the scientific method."