Table of content
1. Performance measurement basics
- Tracking principles
- Site-centric KPIs : how they are built, how they should be interpreted
- Detect traffic source
- Possible biases in measures
- Use segments for a more granular analysis / targeting
- Attribution models
2. Owned : site centric tools
- Overview of most used analytics tools
- Keep track of users’ journey
- Test and learn process : A/B testing
- Customize your key metrics
Tracking principles
Please find hereby an explanation of how tracking works.
Site-centric KPIs : how they are built, how they should be interpreted
You’ll find bellow an explanation of how main site-centric kpis are built.
Events are the cornerstone to accurately track the activity on your website or app, because they allow tailor-made measurement : you keep track of actions specific to your website/app, that nurture your business model.
See hereafter examples of tailor-made measurements you might need to keep track of key actions on your site .
Practice
According to you, what events would correspond to macros and micros conversions on your site/ app ?
Do you think they are all tracked (you can verify this using the browser console, and the ‘network’ module)?
If the answer to the previous question is ‘no’, what do you think is missing ?
Detect traffic source
Why it really matters : knowing the source of traffic allows to determine the associated cost, hence calculate the cost of acquisition or ROI.
An explicite naming convention should be defined and shared among all parties involved in traffic acquisition :
- External agencies (search, display, social, programmatic)
- Internal : CRM department
- If applicable : partners websites / affiliates, … basically every agent helping bring some traffic to your website.
Practice
1 – Just take a look at parameters contained in :
- Promotional emails you receive : there are chances that the links they contain are stuffed with parameters detailing the source, the campaign, and so on. In the case of a tracking with Google Analytics, parameters will start with ‘utm_’.
- Paid search results with Google. Let’s take Renault (French carmaker company) for instance: try and make a first search on Google with the name of the brand (correctly spelled); then do this search again; this time with a spelling mistake. What parameters do you see in the url when you are landed on renault website ?
Finally, enter the name of a vehicle (say ‘clio’ along with the name of the brand). Look at the url parameters again. Have they changed ? If yes, what can you learn from that ?
2 – Build up your url : https://ga-dev-tools.appspot.com/campaign-url-builder/
NB :
i) the url should point to a page of this site (formation.dibenn.com)
ii) you can fill in the utm_source, utm_medium, and utm_campaign fields with any easily recognizable string of character (like your first name), so that you can spot your session in no time on the reporting.
After few minutes, you should be able to see your session on this reporting (click on the square at the bottom right of the report to enlarge it).
Warning : the report will not show any data if you are navigating on private mode.
.
To go further
Here is an example of a file to build up your own urls, using parameters that you’ll have defined and shared with all involved parties.
Possible biases in measures
Legal constraints
Keep in mind that because of legal constraints, you’ll not be able to track the exhaustivity of your website / app traffic. One mandatory prerequisite is the user’s consent to have their navigation tracked. You’ll probably want to use a CMP : Consent management Plateform to collect users’ consent (or not).
As a consequence : if the user’s consent is not collected, you should not fire any tracking tag; which translates in a loss of traffic data. Note that it is possible, though, to implement a minimum tracking under certain conditions (like Google consent mode or ‘hybrid consent’).
.
To go further
Related articles and pages are to be consulted here. password : formationGA
Practice
Check if your site(s) are compliant with GDPR requirements.
Think about it
Pure acquisition campaigns : according to you, what happens when one can consult all the content displayed on a landing page without having to give his consent ? What is the consequence for the advertiser who paid for this campaign ? Is there any workaround you can think of ?
Technical constraints
More and more browsers will just ignore third party cookies. The impact is huge : third party cookies are the backbone of current online advertising system. We are about to witness a dramatic shift in the way advertising compagnies operate.
.
To go further
- https://blog.chromium.org/2020/01/building-more-private-web-path-towards.html
- https://webkit.org/blog/10218/full-third-party-cookie-blocking-and-more/
- https://www.simoahava.com/privacy/intelligent-tracking-prevention-ios-14-ipados-14-safari-14/
The solution can be the implementation of server-to- server tracking systems, what will ‘mislead’ browser, making them believe that the cookies are first party ones. The trick consists in sending the data in a server that has the same name as the domain name of the site where the tracking is performed.
Practice
Take a look at this website : open the browser console (with the ‘network’ module’, and filter out hits that are containing ‘collect’, just to isolate Google Analytics hits. See to what domain they are sent.
Segments
Vilfredo Pareto is an Italian sociologist and economist (1848-1923), famous for having found what would soon be called ‘Pareto’s principle’.
In numerous arars, 20% of actions generate 80% of results. For example, in retail sector, 20% of clients generate 80% of the turnover.
Hence the importance of segments : focus on the most profitable clients allows to secure one’s business, while understanding the type of prospective customers that should be target in the first place.
The key is to understand the behavior that will reflect the belonging to the best-contributors segment. Bellow is an example of a segment of users made up with a selection of users who performed 2 or more sessions and achieved more than one objective within the given time frame (October 1rst to November 30th 2020).
The Recency – Frequency – Amount segmentation is a simple yet efficient basis to isolate within the audience visitors that are the most contributive to your business.
Practice
Take a moment to think of the criteria that you could use to implement this segmentation for your site or app. How many sessions and objective achievements would be appropriate in a one-month period of time ? The answer might vary a lot, depending on the type of business you are considering.
Test and learn process : A/B testing
Practice
Try and find 3 test plans that you think are worth being implemented on your site / app :
- What insights are you trying to grab ? What assumptions would you confirm – or not ?
- What variable(s) would test ?
- On what ground would the ‘champion version’ be selected ?
Attribution Models
Practice
What attribution model do you currently use ? Would it make sense to compare it with another model, in order to take into account specificities like purchase cycle duration or competitive position on a market ?
Overview of most used analytics tools
You’ll find on this page an estimate of the most common tracking tools market share. Please bear in mind that some tools’ usage might look very widespread at first sight, but take the figures with caution. They might be used by a lot of small websites because of they cheap -if any- cost, whereas other ‘major’ paid tools can be used by fewer but bigger companies, best-in-class in their respective areas.
Also, it’s not uncommon to have a website tracked with several tools. It makes sense because there are always discrepancies between one tool and another. The essential thing is to keep the level of those discrepancies constant.
Customize your key metrics
As we saw earlier; the metrics that are the most suited to your app / site is probably everything but standard. Indeed, the simple ‘page view’ with the corresponding url will not be enough to run your analysis.
Very soon, you’ll find yourself wanting to know a lot more about what was displayed on the page / screen view by the user, allowing a much more granular analysis of the users’ journey on your site/app.
This is where custom dimensions come into play. .
Practice
With the help of the adswerve browser extension tool, take a look at custom dimensions collected by Renault on its website through its page view hits; then watch this video illustrating what H&M collects while users broswe its website.
Does it make you think of custom dimensions you would like to collect on your site/app, or maybe on the opposite does it make you more aware of certain type of data that should definitely not be collected, with regards to privacy protection laws ?
