May 17, 2020

How to avoid Analytics Data Overload?

Baiaman @baiaman

Hi everyone,

How do you manage your analytics data and make them useful?

I wanted to know which metrics you track and what are the tools that you use.

Do you track your traffic sources, email subscriptions, etc?

And how much time do you spend on analyzing your analytics data?

And are the insights important enough for you to keep investing your $$ and time on those tools?

We use GA, Seach Console, Facebook Pixel, SERP tracker tool, Bitly, and many more analytics features of the apps we use. Checking each of the analytics platforms every day or even every week doesn't seem to have enough return on our invested time.

Any tips and recommendations would be great!

Thanks!

  1. 4

    I think the less understanding you have of your product, your market and your users - the more you're going to rely on analytics to make decisions. (Plus if you take VC money, you have to have analytics)

    But it is possible to be successful without obsessively tracking every single click people perform on your website.

    I've been doing marketing at Ahrefs since 2015 and we were growing ~50% YoY WITHOUT:

    • Google Analytics;
    • Facebook pixel / retargeting;
    • A/B testing;
    • Conversion tracking;
    • Cohorts;
    • etc.

    Somehow just talking to customers, following our industry and applying some common sense as to what (and how) we should do next always worked for us. We're a ~50 person team that has crossed 40M in ARR back in 2018. Bootstrapped.

    Could we grow even faster if we were to track all those things and include them in our decision making? Who knows? It might have helped us, but it might as well have slowed us down.

    1. 1

      Love this approach 😜

    2. 1

      Great advice here from Tim!

  2. 2

    I know that feeling! I've learned from experience that less is more when talking about data, analytics and other tracking. And this way of thinking helps with page speed, productivity, GDPR and other regulations too.

    Completely agree with the advice from @timsoulo. Truly understanding your audience is more important than any test you can do or hours you spend obsessing about collecting even more data about them. So do focus on understand your audience first and that will help you communicate to them better.

    And then simplify your stack. Choose a handful of main metrics that you really care about and that are insightful and actionable and let those guide your way in the way you analyze your performance, in the marketing action you take etc. It's the action you take that makes a difference rather than looking at stats after all.

    This is one of the reasons we're working on Plausible Analytics. Even website analytics are way overkill collecting hundreds of different data points and making website owners confused and distracted. So we focus on few of the most important metrics and put them all on a simple and easy to understand dashboard that only takes a minute or so to check to understand what's happening. Then you can focus on taking action and making a difference.

  3. 2

    It’s hard to set everything up for analytics tools to work (and I’m working on a tool to automate that with AI!) Here’s the thing about analytics: customers don’t tell what you need, they show. They show if your product is useful, and what parts need to be improved, and what parts you should double down on, using their actions. The numbers aren’t really important themselves, but analyzing user actions is a must for every product IMO. What I suggest is to take your few best users and stalk them through your product. What did they do when they signed up that others didn’t do? Make other users do that. Put that action in the very front. That’s the metric I track.

    Please contact me at my email ([email protected]) or twitter (@bgrgndz) to stay in touch. I can give you a heavily discounted version of our MVP a few weeks later. It integrates to your app with a single line of code and gives you weekly suggestions to improve, no incomprehensible graphs, no numbers that don’t lead you anywhere, just ideas.

  4. 1

    Hey @baiaman, fulltime analytics lead here in a billion-dollar enterprise. (So, my answer to "do you track ____?" is "yes!")

    3 principles I would encourage any business leader to follow:

    1. Start by asking yourself what you want to know. Go to the correct tool for it, get the knowledge, and get on with your day. If you know you don't capture the metric, now you have a very focused bit of analytics work in code to do. If you don't know, ask a professional.

    2. When monitoring a metric, start by deciding what business outcome you want to improve. Decide what the metric is for that. If you want more organic traffic, put Visits from Search (from GA) on a big monitor and then get to it.

    3. Get in, get out, and get back to work. Don't dawdle in your tools.

    It sounds like you may be a little overwhelmed with the options for metrics, interfaces, tools, and so on. You may want my book.

    The goal for the book is to take you from "not sure what to do" to "done" in a few hours, with future-friendly analytics data capture all set up.

    It'll tell you:

    • which tools to use, and why
    • what data to capture, and in prioritized order

    Why me? I've spent the last 8 years in the analytics industry, building scalable, billion-dollar analytics for PlayStation, Rakuten, and a chunk of the Fortune 500.

    Sign up here to be notified, and to receive four quick wins and a quick start cheat sheet as freebies: https://signup.analyticsforindies.com

    And/or, follow the product here on IH (but the freebies come with your email signup):
    https://www.indiehackers.com/product/analytics-for-indies

  5. 1

    I'm happy to look at building a dashboard to pull all that data into 1 place and compare various data points.

    As an example, there's a video on the Lakebed website showing, in real-time, how a dashboard can be created in 3 minutes: https://lakebed.io/in-action

    As another example, I pulled 6 federal government Open Data files together (1,300,000) rows and made them searchable/usable: https://app.lakebed.io/proc_canada

    1. 1

      Thanks for the reply!