End-to-end analytics at the design stage

[ashammi228]

Content
Problem
Solution
Pitfalls
To sum up
In Russia, design is usually understood as something purely visual. While in reality, design is about the process of creation as a whole. Design begins at the moment when the idea of ​​a new product is born. The success of a business directly depends on how high-quality it is.

Problem
Unfortunately, many designers and engineers do not approach design comprehensively simply because the business itself does not require it. Conduct research, design user scenarios, get testing results - and that's it, you can start mocking up and implementing. And only then, at the very end of development, someone from marketing gets the idea of ​​adding analytics to the product. It turns out that it is impossible to track this or that user action within thetelegram customer service malaysia service architecturally, and the poor programmers have to "saw crutches". It is clear what this leads to: product development and technical support become more complicated, UX worsens. Just a couple of such moments can increase the cost of implementing a new feature several times and significantly reduce the interface performance.

To avoid being unfounded, I will give an example.


There was an almost ready service for searching for certain industrial objects. The service had a map with markers that, when the scale changed, were combined into clusters (such circles with numbers). There were a lot of markers, creating clusters turned out to be a very gluttonous process that ate up half the RAM and sometimes completely hung the browser.

They decided to do clustering on the server side, and at the same time moved half of the logic there. And when it turned out that user actions on the map needed to be tracked (and necessarily linked to the acquisition channel), the developers threw up their hands. They had to integrate the analytics service on the server too, generate additional requests, increase the map response time, and so on. Of course, both the architecture and the user experience suffered.

Don't waste your advertising budget
The call tracking and end-to-end analytics service will show the effectiveness of each source in terms of “from click to sale”.

Get a consultation

Solution
Ideally, design should go hand in hand with analytics. But usually, designers (project managers) understand analytics mainly as the analysis of user behavior, carried out at the design stage. In the best case, business analytics is included here. And regular web analytics, for example, is most often forgotten. Not to mention deeper levels such as complex channel and conversion tracking.

At the same time, it is not difficult to receive information from marketers and analysts during the work process about what exactly they will need to track in the new product, and simply include it in the documentation scope. At least in the form of a simple functional diagram:


Developers, having received such data, will be able to design the architecture more competently. Marketers will understand the metrics from the very beginning, making the development of the product more effective. And designers (they are also designers) will receive a more stable and positive user experience.

Pitfalls
However, the world is not perfect. And business even more so. Sometimes at the design stage there is no sign of marketers, and the designer takes on the functions of a business analyst, a designer, and a full-time psychologist. In such a case, there is nowhere to get information about what events need to be monitored. It would seem, nowhere.

But in fact, a normal designer knows about the main technical limitations of platforms, as well as the main target actions of their users. There is nothing complicated in laying down the sending of the main user path to analytics at a basic level, forming an approximate funnel and adding information about the channel for attracting a new user to its very beginning. At least at the diagram level.

To sum up
Of course, everything needs to be done in moderation. It is worth remembering that the cost of an external analytics system most often depends on the number of data sent to it, so sending everything in a row (they say, marketers will figure it out later) is definitely not worth it.

In addition, at the development stage, it is recommended to create a universal system of triggers, to which various actions can be assigned (including the same sending of events to analytics). So, if later you need to track some new indicator, this will not cause any special problems.

Properly configured analytics at the start allows you to very quickly adjust the vector of product development. And if it is integrated flexibly enough, then even a radical change in the vector will not worsen the quality of service. At least, the culprit will definitely not be analytics.