Analyze UX – Ninth Step of UX Strategy From Start To Launch A Product

In my previous articles, I explained the UX first step – UX Research, UX second step – UX Plan, Third step Explore UX, fourth step UX Communication, the Fifth step Create UX, Sixth step – UX Feedback, Seventh step – Finalise UX and Eighth Step – UX Delight. In this article, I discussed how to Analyze UX with Detailed KPI’s Setup, including your goals, achievements and Write down how you define success and failure. A/B Testing planning will help to improve your KPIs with User surveys, sessions recording, observe and fix, test, observe and fix.

Let’s read in detail below:-

Web designing is all about striving for the best. This can be seen in the data-driven design as well where the data of the product is compared in different interfaces so as to ascertain the competence. Good metrics are required to track the progress and see where the product is going.

These selected metrics are used to learn about the quality of user experience, also known as the HEART framework and to know the goals of the project that is known as the Goals-Signals-Metrics process. The HEART framework included happiness that looks into the attitudes or the behavior, engagement that sees how much the user was involved through the frequency or depth of interaction.

Further, it has adoption and retention that judge the new users and the existing users that are returning respectively. Lastly, there is task success that takes into account traditional behavioral metrics relating to user experience. The HEART framework helps in selective approach and decides upon the categories that are the most important.

In the Goals-Signals-Metrics process, you need to identify the goals so that you can choose the metrics accordingly and then measure the progress towards these targets. A limitation that can be faced here is to fix the goals as per the existing metrics. Thus, proper dimension has to be determined with the consent of all the team members.

Next, comes signals which are to track your goals. Again, you will have a huge number of signals associated with a particular goal. Research will eventually help you to cut down and choose the necessary. Select those signals that are sensitive to design and are the best determinants of the relative goal. Signals are further improved upon into metrics that are ultimately used for the A/B test. Select the most appropriate metrics through analysis and data research. But most importantly track the metrics corresponding to the top goals as it will also reflect the quality of the user experience.

A/B testing is kind of optimization technique in which the conversion rates of two different versions of a page, namely A and B are compared by way of live traffic so as to assess which page is most effective. This can be done through the buttons clicked or video watched or through the signing up of newsletters in these versions.

A/B tests can be used to determine the testing of two completely different design directions. Now, for comparison, the entire design is taken into consideration rather than the individual elements as the target is to track the page’s business goal. In case of debate for one element, this type of testing is very useful by way of splitting the percentage determined.

A/B testing is a competent method of testing and is preferred the most. Make sure the tracked variables are limited, so that the tests can give reliable results without having to monitor a large amount of traffic. If the website has small number of visitors and further it is split into three or four segments it will be really hard to finish the test.

A/B being a speedy way of testing is used by a large number of sites for the purpose of testing and is preferred more over the multivariate tests. A/B testing is a versatile way of optimization and should not be complicated by using more variables. Let the best page win!

One effective way of testing is pilot testing that takes into consideration practice sessions before the actual scheduled study takes place. These sessions help to prepare for the full study and makes sure that everything is in order. Pilot testing is a good option if you are new to the whole process of testing as you can disregard the sessions easily. It can also be a good move if the testing is conducted in the unfamiliar zone as it will help to prepare you if you are not an expert. In this method, all the communications be it from recruiting to, confirmation emails need to be tested.

This is done to limit the possibility of misinterpretation and also so that the instructions can stand by itself without having to give any answers. In the case of high visibility project, pilot testing is the obvious choice as the project demands extra care. It is kind of an extra insurance given to certify smooth running.

Dress Rehearsal- It is just like a dress rehearsal where the dry run is done to insure the final preparation for the study. It creates a low-pressure setting that makes sure the team is prepared and set for the final test run.

Test the Tasks – It is important to find and fix the tasks before proceeding on to the final study so as to get more accurate results. This way confusion and misleading can be avoided from the task level itself.

Timing – It is difficult to determine the time required for testing. Pilot testing gives you a rough time estimate thus helping you to prioritize your tasks effectively.

To get the most out of the pilot testing, make sure to schedule the test one day in advance so that you can pilot instructions for a longer period of time. Recruit the people who have the similar profile as that of target studies. This helps in getting valuable feedback about the site or the study. Pilot study does have its challenges in terms of materials preparation, extra participants and time needs, but the payoff will be better as it is through the smooth final study we will be able to derive the results and the product’s success rate.

http://www.gv.com/lib/how-to-choose-the-right-ux-metrics-for-your-product

http://www.dtelepathy.com/ux-metrics/

http://conversionxl.com/8-effective-ways-of-measuring-ux/

http://www.measuringu.com/blog/ab-testing.php

http://www.nngroup.com/articles/which-ux-research-methods/