Leave a comment
Get the GH Bookmarklet

Ask GH

Can you describe the analytics stack? Do you write your own queries? Do you use custom analytics? How often do you doubt your data? I'm very curious about these questions, I <3 writing my own queries and planning the event schemes, but I'm starting to have second thoughts about it :)

  • TH

    Tristan Handy

    over 5 years ago #

    Hi Dan-ya! I run the 12-person marketing team at RJMetrics and have a background in analytics. We use our own product, RJMetrics Pipeline, to consolidate all of this data into an 8-node Redshift cluster:

    - product usage (mysql)
    - Zuora (payments)
    - Snowplow (web analytics)
    - Pardot (email / marketing automation)
    - Zendesk (customer support)
    - Salesforce (CRM)
    - Many internal services / data sources

    In all, we probably have 15-20 distinct data sources being piped into Redshift on an ongoing basis.

    Here's what I think is so important about your question though: I spend a meaningful amount of my time analyzing this data writing SQL. Redshift is amazingly performant, and it makes it very fast for someone who is familiar with the data and an expert SQL user to get answers that previously just weren't possible. I use a combination of SQL Workbench to actually query Redshift and then Mode Analytics to visualize this data and share it.

    I find that it's incredibly important for me to do my own analytics, because it's in this process where my most creative thinking happens. I can actually see what is going on with our users in a way that I just can't when I ask someone else to do the analysis, and I end up having insights that I just wouldn't otherwise. I don't know if this is just my personal style or whether it is objectively a good strategy for how to run a marketing team, but it's what works for me.

    • DS

      Dan-ya Shwartz Bar-El

      over 5 years ago #

      Hi Tristian,
      I really appreciate your answer.
      thank you for taking the time.
      (I know this may sound like a generic answer, it isn't)

  • MB

    Morgan Brown

    over 5 years ago #

    Hey Dan-ya,

    Here's what our analytics stack looks like where I'm at. It's not the most complicated in the world, and we've still got a long way to go, but this is what it looks like today:

    MySQL user DB: tracks all user data, actions, etc.
    Stripe: tracks all payment and customer data
    Segment: tracks all event activity on the website, identity, 3rd party tools integration/etc.
    Google Analytics: website traffic data
    Chart beat: realtime traffic data (We're a media publisher, fwiw)
    ChartMogul: recurring revenue tracking, etc.

    What we're currently building:
    Warehouse fed by: internal database data, Segment events, Stripe data, email data

    What we're going to next: something like a Looker to empower the team to find deeper insights.

    Right now, most of our user-level analysis is done by our data scientist, she was one of the very first hires I made at Inman. My second hire, our growth lead, Kyle, has an accounting background and used to do auditing for a big accounting firm here in the US.

    They're both very analytical, so they do the number crunching for our team. I do qualitative analysis and surface-level web/content analysis, and work with them on user surveys, our testing plan, etc.

  • CB

    Charlie Benkendorf

    over 5 years ago #

    I'd say it depends on stage of your company. As you scale, you can afford to have dedicated business intelligence teams that takes a lot of burden off. But I'd say you always will write your own queries / create your own reports, sense check your data, etc. At the end of the day it's for your benefit so if something in the existing system isn't working for you, you'd need to point it out.

    Our stack is like this:
    -SQL holds a ton of info on visits, leads, loans, and even acquisition costs for the loans. That took a while to get right...
    -Google analytics / mixpanel for website traffic
    -Looker for building SQL queries / reports

    I'm sure there's a ton of other systems but those are the main ones I touch in remarketing.

  • PI

    Petru Iacob

    over 5 years ago #

    You know what they say: nobody ever complained they've tracked too much. Most startups under track their business' evolution. At Neonmetrics, we use Google Analytics to record traffic activity in the public pages and our custom reports to view it. Once logged in, all users are being tracked so we know when someone creates new reports, connects new platforms, etc. We do all of this internal tracking using MongoDB. Basically, same thing as Keen.io or Mixpanel, just using our own machines as it's faster.

    Like Charlie said above, it really depends on the stage of your company. In the beginning, out of the box tools/reports might be just fine, but as you go, you'll notice you need more advanced queries.

    Hope it helps!

  • SD

    Santu dey

    over 5 years ago #

    We are using

    1. Mysql + looker for user DB
    2. Event tracking - Mixpanel
    3. Session tracking - Google
    4. Mandril and Sendy for email marketing (Click+ open rate)
    5. Notification - Own tool + Mixpanel
    6. branch.io - Deep linking + acquisition source + content analytics + Referrals tracking

    • DS

      Dan-ya Shwartz Bar-El

      over 5 years ago #

      no one dashboard to rule them all? how do you join you growth campaigns with your email and notification campaigns?

      • SD

        Santu dey

        over 5 years ago #

        We are on the way of building a dashboard to track all the campaign for now. Joining email and notification is a very difficult task after trying some software we realize that it will be great to have our own software.

      • SD

        Santu dey

        over 5 years ago #

        We are on the way of building a dashboard to track all the campaign for now. Joining email and notification is a very difficult task after trying some software we realize that it will be great to have our own software. So we made our own synchronize system to send email and notification.

  • AH

    Agnes Haryuni

    over 5 years ago #

    I use Google Analytics, various data provider that provide me statistics of the market trends, social media management tools, and in-house tracking system for user retention and traffic

  • JP

    Jesslyn Phoon

    over 5 years ago #

    Hi Dan-ya,

    I'm no expert but allow me to give you my two cents' worth. :)

    I'm currently the sole person handling the data analytics in my company as we're a small team of 7.

    I guess there is never a clear definition of getting "too deep in the data". In my case, there's always so much more depth I can cover but it really depends on the resources (time, effort, money) that is available as well.

    As we started off with very limited resources, here's what I have chosen to use in the past year. Most of them are free - the ones which i have paid for usually serve dual purposes (I will explain more below):

    1. SQL for tracking in-app data
    2. Google Analytics, Mixpanel, and Search Console to identify website traffic data
    3. HotJar for slightly more in-depth website user behavior (funneling, heat map, polls...)
    4. Intercom (doubles as both an automated customer support and basic CRM platform)
    5. InlineManual (doubles as an onboarding and in-app user tracking tool)
    6. Yoast Analytics and Buffer for content analytics

    I'm planning to check out Looker this year.

    I hope this helps.