What is the difference between EYE tracking and MOUSE tracking?

Over the past few years, the usability market has been actively using such a tool, no doubt a convenient and good one, like mouse tracking - a system that allows you to track the visitor’s mouse movement on your site and the clicks made by it. The main player in the Russian market in this area is Yandex.Metrica , which bought a web browser that quickly gained popularity.

This tool is very convenient: it allows you to build an activity map on the page, highlighting in hot colors those areas where the mouse moved more actively and more often, a click map showing which places on the page were clicked more often. All this information can be a good source of information to identify the amenities and disadvantages of your site.

And it’s quite natural that the studios, usabilityists and site owners who are not familiar with IT tracking have a question: how are the maps built using the IT tracking data different from metric maps and which system is better.

I must say right away that neither system is better or worse - they are simply different. The metric is more convenient for some tasks, aitracking is for others, and if used correctly, both can be quite successfully used to analyze the site’s convenience. But let's take a closer look.

Heatmaps

It must be understood that a “heat map” is nothing more than a method of displaying information. Just in one case, information on the intensity of mouse movement is superimposed on the screenshot of the page, and in the other, on the movement of the gaze. Those. in principle, these two cards are not connected at all with each other. With the same success, it is possible, for example, to impose data on the speed of reading a text on a page and also get a heat map.

Another thing is that there is a definite correlation between the movement of the gaze and the movement of the mouse. But this correlation is completely ambiguous and is not manifested by all users.

Correlation between gaze and mouse movement

I know only one scientific (read, fairly objective) work that explored this issue: What can a mouse cursor tell us more ?: correlation of eye / mouse movements on web browsing . It says about the strict connection of the mouse with the eyes - 84-88%. (I would be grateful for links to other similar studies) But this article has two drawbacks.

First: it states that with a probability of 84%, if there was a cursor at some place on the page, the site visitor looked there (not necessarily at the same time!), And with a probability of 88%, if at some I didn’t look at the place, then the cursor was not there either.

Of the regions that a mouse cursor visited, 84% of them were also visited by an eye gaze. Furthermore, among the regions that the eye gaze didn't visit, 88% of them were not visited by the mouse cursor, either.

But this does not at all follow that the cursor was everywhere the visitor looked! Those. a heatmap of mouse movements shows a certain (unknown how big or small) part of what the user supposedly saw. The second conclusion is that it is impossible to evaluate temporal parameters of information perception by mouse movement, including the sequence of this perception.

Try to keep an eye on yourself while visiting sites and keep track of exactly what you see and how you move the mouse.

The second:this study was conducted in 2000, when sites were not as widespread as they are now, and users were less familiar with computers. And my observations (not yet framed as a scientific article due to insufficient statistics) show that the correlation of gaze and mouse decreases with increasing user experience.

At least during our research, in 6 out of 10 cases, experienced Internet users do not move the mouse at all until they find what they want to click or check for hover (the cursor remains almost in place, and the mouse wheel is used to scroll the page ); among inexperienced users, this behavior is observed on average in 3 cases out of 10. Cases of a more or less accurate cursor following the gaze (coincidence of the position of the cursor and the trajectory of the gaze at least once per second and a half) are much less common, approximately 1 out of 30. Also approximately every tenth user moves the mouse absolutely not where his gaze is (presumably, a habit left over from the old times to prevent the screensaver from starting up).

To some extent, this observation of mine coincides with the fact that in later studies (not so scientific, and involving a smaller number of respondents), “percent correlation” turned out to be lower: 69% in the 2005 study , 42% in the 2007 Google study ( p. 31), and 32% in a 2008 Google study , where mouse movements were more or less similar to gaze movement in 56 cases out of 175.

Conclusion: there is a definite correlation between cursor and gaze movement, but how significant is it and in what cases it is applicable - not quite clear.

Conscious and Unconscious

The reason for the discrepancy described above and the main difference between eye tracking and other research methods is that it captures unconscious reactions, and mouse movements and other actions are conscious. If you simplify it a bit, then with the help of aitracking you can track (and often just see) the reason for the action, and with traditional methods you see the result and, analyzing it, try to get to the bottom of the reason.

By the way, it is thanks to the fixation of unconscious reactions that eye tracking allows you to reduce the required number of respondents in the focus group. If for traditional methods the minimum number is 8-10 respondents, then for aitracking it is 5 respondents for qualitative (as opposed to quantitative) analysis (optimally 10-15).

Here it will be appropriate to additionally provide a link to a good article by Jacob Nielsen " Five users - everything you need for the test "

Number of respondents

Of course, a huge positive factor of the webvisor is the large number of visitors to your site for which statistics are collected. Moreover, these are real visitors, and not some way of a selected focus group. Such data is of great value, allowing you to evaluate the flow of visitors to your site and their activity on its pages.

Respondent Objectives

The plus of the webvisor described above - a large number of real visitors to the site - paradoxically turns into its own minus: you do not know why all these people came to the site and far from always know who they are. That is, in order to evaluate, for example, the convenience of making a purchase in a store, you need to somehow separate the visitors who came to buy something (including those who could not do it) from those who came, say, simply see what’s new or find out the characteristics of the product.

When testing on focus groups, you know which people perform the task, which task they perform and have the opportunity to conduct a survey during or after the experiment and ask them to explain some of their actions. This is very useful from the point of view of identifying and eliminating various usability errors.

We can give such an analogy. The web browser is information about the movement of cars around the city. The most different cars, with different drivers: someone just rides, someone goes to the store, someone to the cottage, someone to work, and someone broke. This information is very useful from the point of view of the urban economy: you can see where traffic jams occur and try to eliminate them.

Aitracking in this analogy is “launching” several “typical” drivers with a proposal to get from point A to point B and thoroughly fixing everything that is happening. These data allow us to identify the maximum inconvenience that prevents drivers from reaching the coveted point B and, accordingly, to make this route the most convenient and understandable.

Price

High cost has traditionally been the main disadvantage of IT tracking - expensive equipment, highly professional employees, the need to recruit and pay groups of respondents - all this requires money. However, the research price is by no means exorbitant - most studies cost about 20 thousand rubles, and simple ones start at 6000. On our website you can estimate the cost of the study on a calculator and see what it is made of and what you can save on.

Research materials

In terms of what can be investigated using these methods, eye-tracking wins. If mouse tracking allows you to explore only sites and, with certain restrictions, program interfaces, eye-tracking, in addition, is used to study design layouts and prototypes, printing, commercials and other materials that may not require any action from the user.

Moreover, one of the most effective methods of using IT tracking is just testing layouts and prototypes - you get the opportunity to identify errors at the earliest stages when the cost of fixing them is minimal.

Conclusion

Both mouse tracking and eye-tracking are quite convenient and useful tools, and contrasting them with each other is pretty stupid. It is unlikely that you will oppose the hammer with a screwdriver or microwave to a coffee maker. You just need to understand what you want to achieve and what parameters to measure.