Four Steps to a Free Data Analysis Platform for Your Product "Product people - Product managers, product designers, UX designers, UX researchers, Business analysts, developers, makers & entrepreneurs December 12 2014 True Analytics, Data, Data Analysis, Usability Testing, Mind the Product Mind the Product Ltd 1331 An example dashing dashboard Product Management 5.324

Four Steps to a Free Data Analysis Platform for Your Product

BY ON

Understanding how customers use your product is a vital part of product management. To get this insight of how customers are interacting with your product, there are two different routes open to you: Qualitative research, or Quantitative research.

Qualitative research involves such methods as user studies, sitting with a focus group, running workshops with users, asking them to fill out questionnaires and observe how they use your product, and generally asking them for feedback. This is a very valuable part of product research and shouldn’t be ignored, as it offers highly rich insight.

Quantitative research on the other hand involves looking at data generated from your digital product. This is means recording and analysing various metrics from your site to measure the performance of your product over time and to track changes in user behaviour. This type of information is equally crucial to product research, as it allows you to analyse the activity of all your users, not just a small focus group. But it’s generally the most complicated to carry out.

Product Managers, Data Analysts and company founders often believe that they need to purchase expensive Business Intelligence software to feasibly collect and understand all of this data. However, that’s not always necessary. In this article I share 4 steps that could help you build a free analytics platform for your product.

Why would you want to build a data analysis system?

There are a large number of business intelligence software and insight tools available for businesses. These software tools are expensive to implement and maintain and often require specialised training to use.

Building your own data analysis software will probably take the same amount of time as you would spend learning one of these tools, but you can use standard programming languages such as Python and Javascript, and avoid the need to learn any one vendor’s proprietary code. Creating your own analytics system also gives you 100% control of your data and what happens to it.

The four steps:

Step 1: Automate Google Analytics

You need to have Google Analytics installed and running without problems before you start (think of that as step zero). The first real step then is to automate the recordings of your KPIs (key performance indicator) in a Google Doc. This can be done using Google’s Analytics Report Automation (Magic script). You can find out what the script does by watching Nick Mihailovski’s video:

This script allows you to generate reports on data points of interest, that relate to your KPI’s. To find the exact name for the points of interest in the Google API use the Google API support tool also mentioned in the video.

Once you have the data you want coming in to your Google Doc you can use  Google spreadsheet functions to gain deeper insight. This can be automated too using the vlookup and index functions in the spreadsheet. I created a sample spreadsheet for this article, so that ou can see what this looks like here on the results tab: here.

The function I use calculates the percentage of visits to my site that originate on mobile/tablet devices, and on desktop. The calculation function I use is in cell B2 on the tab Results:


if(len(A2)>1,round((INDEX('Device Day'!A16:C51,MATCH(1,('Device Day'!A16:A51=A2)*('Device Day'!B16:B51=$B$1),0),3)/vlookup(A2,'Sessions Day'!A16:B27,2))*100,2),"")

 

More recently, after the Mihailovski video was released, and after I started using the automation script, Google released a plugin for their spreadsheet, connecting it to Google Analytics. In principle everything is the same, however you can’t automate the script from the plugin.. yet.

Google Spreadsheet Plugin for Google Analytics
Google Spreadsheet Plugin for Google Analytics

So what does this all mean for your business? Making use of these free Google tools allows you to file your product data directly into a spreadsheet and split this into dimensions of your choice. Furthermore, once the data is in your spreadsheet you or your analyst can run standard functions to gain even more specific and granular insight.

2) Use scripts to automate social media recordings

Google Analytics isn’t the only potential source of data for your analytics system. There are a number of social media scripts that you can direct to Google Spreadsheets too. These include scripts to collect data on Twitter followers, Facebook Page likes and Foursquare check-ins such as those developed by by Martin Hassman, http://labs.met.cz/. Using Martin’s scripts, you are able to pull in data from your social media account into your spreadsheet, and from there calculate changes in social following over time.

How does Sheetsee.js work
How does Sheetsee.js work

3) Use If This Then That (IFTTT) to automate social media and other web event recordings

To get even more detail of your products’ and brand’s social media impact, you can use IFTTT to automatically record messages on Twitter using a hashtag or mention, in to a Google spreadsheet: An example is on IFTTT. Again, once this data is collected, you can run calculations such as daily engagement to look at the number of tweets you receive per day.

4) Use free online dashboarding and graphing tool for your Spreadsheets

In Nick Mihailovski’s video he suggests presenting the final results from your report using a Google chart and default webpage.

I would also recommend improving on this by embedding these charts on your own HTML webpage. You can take this a step further by creating your own dashboard using free web tools such as sheetsee.js, dashing.io or charts.js, for example.

I have used sheetsee.js, by Jessica Lord, @jlord as it is built to integrate directly with Google Spreadsheets.

Sheetsee.js is built on top of a library called tabletop.js, a library which is built to expose spreadsheet data in javascript.

This is a fantastic way to expose spreadsheet data in a smart, graphical manner. The documentation for both libraries is clearly detailed, and it does not take a lot of coding skill to set up and start working with this tool.

Even if it means you or your analysis team need to learn some Javascript, unlike purchased analytics systems this is at least a transferable skill that does not keep you locked to a single tool.

Dashing is a dashboarding tool built by Shopify, built on Ruby. It too has a relatively well documented library. It did take me a while to set up, but I felt the results were worth the struggle.

An example dashing dashboard
An example dashing dashboard

Chartjs.org is another free javascript charting tool you could consider using.

I hope I’ve been able to demonstrate how developing an effective analytics system is an accessible option that doesn’t always mean working with expensive proprietary software. There are options for developing your own analytics system, using cheaper, standard tools, and cheaper (or free) scripts. You will maintain control over your data, create the analysis you need specifically for your product, and you (or your analyst) may also gain transferable coding skills.

And not to forget the caveat: when does this method not work?

This article focuses on web applications and how to use Google analytics alongside other data sources to create and analytics reporting tools. You can use these techniques for applications that have up to 500,000 sessions per month. However it’s important to bear in mind that above this number of sessions Google only shows sample data, which makes the data results difficult to verify. For more details on this problem see Google Analytic Data Support Page. Similiarly, these techniques are intended for web applications only, and you won’t be able to use them for apps that run on mobile devices or static desktops.

I would love to hear your experiences of either proprietary or home-grown analytics systems, as well as any free-tools in the comments below.