My recommendation is to stop all recommendations in design.

It is easy to trust customers when they ask us designers to give them recommendations. We believe that they love us for wisdom, knowledge and experience. They are waiting for our advice. And we love to advise them. They want to do the right thing and we know how to help. And really, what's wrong with our recommendations? They want to receive them. We want to give them. Why shouldn't we do this?

It's simple: the recommendations do not work. The result looks bad. Customers lose faith in our abilities. And the design is not getting any better. Our research shows that the best development teams do not give recommendations. Instead, they use an experimental approach.
')

Patient bending-straightening the arm: "Doctor, doctor! It hurts when I do that."
Doctor, examining the patient: "Hmm. Okay, as a doctor, I recommend you not to do that."

Easy way

Making recommendations is the easy way. You say, "Do this. Change it." and wash your hands. If they don’t, they’re idiots. Ideally, they should follow your recommendations and improve their design. However, this is only one of four possible developments.

They follow the guidelines and the design gets better.	They do not follow the recommendations and the design gets better.
They follow the recommendations, but the design situation is not getting better, or is getting worse.	They do not follow the recommendations and the situation with the design does not get better, or deteriorates

What happens if they follow our recommendations and it does not help to improve the design? What happens if they do not follow the recommendations - they will do everything in their own way, and their design will become better? In any of these cases, it will be difficult for you to continue working with this client.

Making changes requires resources. If there are no improvements, it means that the company wasted energy, time and money. Have we thought about such a possible outcome when we made recommendations?

Let's play “Put your wages on the line”?

Megan Food (specialist in the field of interaction and usability research (Symantec, PayPal (eBay), Wells Fargo Bank)) has one simple rule for himself and his team. An employee can give a recommendation only if he is ready to put his annual salary on the fact that the design of the client will be better.

Can you act like Megan, giving the next batch of their recommendations? Take the checkbook, specify the amount of your annual income net of taxes in the check, and send the check to the customer with a note that he can cash it if your recommendations do not work. How confident are you now in your own recommendations?

Experimental approach

After conducting a study, we found out that a typical recommendation is as follows: " Users experience difficulties because they do not see the names of the form fields, I recommend placing the names over the fields, not the left. "

However, some teams use a different approach: " It seems to us that users have difficulty with the field names. Let's do an experiment, put the names over the fields, not the left, and see if this improves the situation ."

It would seem that the difference is small. However, this approach was demonstrated by professionals with a solid reputation and the best indicators in their field. They told us that they never give recommendations, but they never refuse to experiment.

Discussion of research results

I find the process of work of such highly-qualified teams very interesting. It begins with a general discussion of the source data and their meanings. The team begins to consider all possible sources of the problem. "Maybe users do not notice the name because they are moved too far? Maybe the font is difficult to read? Maybe they did not understand the conditions? We did not correctly set the tasks for the study?" .

Then the team moves from talking to a new study that answers all these “can?”, Reduces the number of possible answers and, as a result, helps to confirm that the user has begun to better understand the interface. Often, the case is not limited to one stage - the results are discussed and the research is conducted again.

This process is significantly different from the method of "recommendations", when one expert gives an assessment of the design and makes up a certain final list of elements that need to be changed. Instead of putting all the responsibility for making a decision on one specialist, making the whole package of changes to the design and waiting for the result, our team has the opportunity to move the design to success in small steps, from experiment to experiment. Moving in steps, you can clearly say which changes work and which do not. Such an approach quickly teaches team members and gives them a true vision of the whole picture, otherwise they would have to follow the “tenets” of the theorists all their lives.

As a result, the team becomes better informed about all the nuances of the design they are building. No one is ultimately responsible for the design, no one is obliged to know everything and be always right, so all team members feel comfortable. Changes made are not viewed as final, but as an ongoing process of improvement.

Change mentality

Making yourself reject recommendations is very difficult. As I said, recommendations are the easy way, so he wants to seem like the best way. But ultimately it's a trap. The probability of a disastrous outcome is too great.

Both experience and research tell us that experimentation, when one constant changes, and the result is measured, serves the team as a good guide to action and, ultimately, leads to long-term success and better results.

This is my recommendation and I insist on it.

Translation from English. Posted by: Jared Spool. Original here