30 days of Ecommerce Testing Day 8: Analytics

The list doesn’t tell us to post here, but i think it would be interesting to share and see any other answers.

  • Know how the user reached the webpage (e.g. google, advertising, direct access…)
  • Know how the user searches for something (e.g sidebar, search bar…)
  • Be able to create a profile for a given user. What does interest him/her most.

Ok @andrepm, you’ve started the business side so I’ll start with the product side:

  • for importing products/items - import success and data consistency monitoring
  • server status monitoring (natural for AWS/Azure but not nearly as much elsewhere) - I’ve seen lack of monitoring lead to disrupting the project that included load-balancing
  • performance monitoring to for early recognition of UX bottlenecks

Updated my bit here https://wp.me/p9EXXo-3l

Talked mostly about monitoring load, performance, and localization.


There are some tough acts to follow in this thread, not the least of which is @ceedubsnz very well done blog post.

This article talks about the idea of “funnels” which is tracking the step by step process of seeing a page, clicking a post, adding to cart, checking out, and finalizing the process. Where in the “funnel” might people be dropping out?

There are other examples, such as email campaigns.
email sent --> email successfully delivered --> email opened --> link clicked in email --> item added to cart --> check out

Google analytics, mentioned above by Mike also does funnel analytics and visulization:

Now, for the ecommerce business, I suppose the goal i s to maximize the amount of successful transactions relative to the initial point of contact in order to make the most $$.

For a tester though? I think we are interested in the same analytics, and the end result is the same (to make the customer money!). However, if we are witnessing a large drop out rate at a certain point in the process, we need to make sure this isn’t due to some technical issue that can be resolved. It would be awful to lost purchases just because of a software bug. That’d be like having to leave a retail store on account of a broken register.

I have left retail stores on account of log lines. I suppose the online equivalent would simply be the amount of time it would take to check something out, i.e. performance. So we’d want to know that too.

That’s my thoughts concerning something I literally learned about 10 minutes ago.

-Dave K


Slightly late to the party on this one and still a mountain of material to go through (apparently there may be almost 70 commercial Key Performance Indicators…), but my intuition would be to go along the lines of relevant quality characteristics.
Mind you, this is Web-only and focused on regular shopping from the buyer point of view, so it’s just the tip of the iceberg. After all, there are also mobile solutions and more specialized commercial services.


  • page load times
  • search times
  • under normal and high load, under stress


  • failover monitoring
  • checkout errors and data consistency


  • certificate problems
  • access to payment information
  • access to customer data


  • the paths to product pages the users take including the number of clicks
  • the time spent in various stages of the checkout,
  • browse/filter vs. search behaviours
  • most popular search terms and browsing targets in the help section
  • localization


  • monitoring the state of interfaces to payment gateways and possibly other systems such as package tracking or social media shopping, if available


  • most and least used features
  • most and least used filter criteria
  • most and least popular payment and shipping methods
  • all the inventory and order status tracking from the seller’s perspective would probably fall under this category but this is again a huge field


  • most and least popular categories
  • most popular search terms
  • most used parts of product descriptions (where it can be discerned)
  • search terms for which there were no results (maybe together with subsequently used terms)

User behaviours:

  • client software used to visit the website
  • peak and nadir time of visits
  • locations from where users visit (both geographical and referrers)
  • registrations from new visitors, if applicable
  • repeated visits
  • conversion from various ways the products are presented including marketing campaigns and upsell features
  • conversion from add-to-cart to checkout
  • reasons for cancelling orders
  • problems with payment and delivery (e.g. packages undeliverable at the target address and not picked up at a designated backup location)

And this is just the beginning. So many sources for test ideas!


Day 8 - Monitoring and analytics

Does this come in testing bucket, just got curious as i had no idea of monitoring and analytics of e-commerce website

I googled and read, found that monitoring and analytics has to do with marketing and promotion of website and thus we could provide a feedback saying how the website could be improvised

Wanted to check if my understanding is right

1 Like

Wow, look at all the amazing answers here! I feel like mine is very lightweight now (no research, just a stream of consciousness), but here you go anyway. https://www.supertestingbros.com/blog/2018/5/12/30-days-of-e-commerce-testing-day-eight

1 Like

Monitoring and analytics can for example help prioritize testing as you have an idea on which areas and elements are used most often or most intensely. On the other hand, when others are not used, maybe they need to be reworked/improved, like you said @divya.

Oh thanks for providing an insight on the same @maos

From the Twittersphere:


and a recent update from @yogitakl

1 Like