Many Digital & Web Analytics positions simply look for an understanding of Google Analytics, Adobe Analytics, Google Tag Manager, etc.
However, what should you be focusing on if you want to really get to the next level? How do you make yourself stand out, and your value much higher? How do you become tool agnostic so that you’re more flexible in a variety of situations?
These are questions I am continually asking myself in order to make sure I’m on the correct path for where I am trying to get to. They’re healthy, and really can help you put things into perspective. After reviewing, here are the 4 technical skills I find myself using most often:
Scenario 1: Personalizing A Unique Welcome Message Based On Traffic Sources.
Scenario 2: Return User Detection For Vendor Tags
I recently had a scenario where I needed to set specific vendor tags to only fire for return visitors. Since the site has no authentication, it’s difficult to tell if a user is return or not unless we are writing some ID to the database. Instead what I did was cookie a user on their first visit each time. This is not perfect since cookies are isolated to single browsers on a single device (and can be wiped), but this is a good start in a non-multi device situation.
Everytime a user comes back to the site I check for this cookie. If it exists I trigger the required vendor tags.
Skill #2: SQL
I have labeled this broadly because there are various forms such as MySQL and PostgresSQL. My suggestion here is to pick one to start with and wait till you need the other. If you will be working in Amazon RedShift, PostgresSQL may be the way to go.
Familiarize yourself with basic select statements, joining, group bys, ordering, window functions, etc. SQL is absolutely an invaluable skill for any analyst. It allows you to manipulate raw or aggregated data very quickly.
Scenario 1: Connecting the dots between Web, Call, and Walk-ins
I had a scenario for an e-commerce client who also had a call center and a few walk in store fronts. They wanted to know how many people who started at each touch point converted at others. They also wanted to know of these people who placed >= 2 orders, how often were they placed at the same touch point as the previous converting touch point.
They were trying to get customers who converted at store fronts to convert at the web later down the road as their profit margin on web orders were higher.
Skill #3: A scripting language (Python, PHP, etc)
The ability to write scripts has made my life much easier. There are cases where I sometimes want to run web data through an API or collect data from an API to use when evaluating something. These are all cases where being able to write in either PHP or Python has been helpful.
I personally started with PHP, however I have been working with Python much more recently. It is better suited for data analysis long term in my opinion.
Scenario 1: Speech To Text transcription for calls that abandoned a web checkout page
I recently did an analysis that showed web visitors who were filling out check out forms with valid inputs were calling instead of finishing their purchase on the web. I wanted to understand why they were calling, and if there was anything we could do to reduce these types of calls.
Fortunately I have access to the raw data here, but I was able to tell users who called after starting a valid checkout, grab their recordings, and run them through a speech to text API, and get transcripts.
This would not have been possible all on my own had I not known a scripting language.
Skill #4: Basic Statistics
A big portion of Web Analytics is gathering data to make informed experiments. How can you make informed experiments (read: test variations) if you can’t confirm you’re looking at a big enough sample size?
Basic statistical knowledge would include things like P-Values, Z-Score, Power, Confidence Intervals, Sample Size Calculation, Statistical Significance (1 vs 2 tail t-test), Chi Squared test, Equivalence testing, etc. Recently Bayesian statistics have been coming into the web picture more and more. Make sure you’re aware of what’s going on in the ecosystem around you, even if it’s not the statistical methodology you’re currently working with.
Latest posts by austin (see all)
- Google Tag Manager: Image Click Tracking - August 19, 2017
- 4 Technical Skills Any Digital/Web Analyst Should Have - August 19, 2017
- 3 Google Analytics Filters For Data Cleaning - March 30, 2017