What is "Persona"? Is this the user portrait used by operations?

The term " user portrait " is actually most likely used incorrectly on the domestic Internet.

There are two English words here. The first one is "PERSONA", which is a typical user model proposed by Allen Cooper, obtained through surveys and questionnaires, and used for product demand exploration and interaction design.

in:

P stands for Primary: whether the user persona is based on situational interviews with real users;

E stands for empathy: refers to whether the user persona, which includes the name, photo, and product-related description, is empathetic;

R stands for Realistic: Does the persona seem like a real person to the people who interact with the customer every day?

S stands for singularity: whether each user is unique and has few similarities with each other;

O stands for Objectives: Does the user role contain high-level goals related to the product and does it contain keywords to describe the goals?

N stands for Number: Is the number of personas small enough so that the design team can remember the name of each persona, as well as one primary persona?

A stands for Applicable: Can the design team use the user persona as a practical tool to make design decisions?

Another word, "Profile", is a data analysis method that uses the data already obtained to outline user needs and preferences.

Both words can be translated as "user portrait", but the first one is used for product research and interaction design, and the second one is used for operations and data analysis.

Therefore, we must first clarify which one to use. Considering Liang Ge’s user structure, I will talk about the second one.

How to obtain user portraits

The basis for obtaining user portraits is data.

This is nonsense, but for some startups , it is the truth.

If we were to analyze a person today, from which angles would we analyze him? Let's recall what we were asked to fill out in the forms we have filled out over the years:

1. Name
2. gender
3. nationality
4. Place of residence
5. Current residence
6. Contact Details

The above seems to be the standard content to be filled in, but you will also encounter:

• blood type
• constellation
• Hobbies
• Work Experience
• Family members
• Income

Even more, let's stop here.

Usually, if we want to understand a person, we need to know his or her basic information, which is the parts 1 to 6 above. Of course, it can be more or less depending on the needs, but generally it will not exceed the parts listed above.

In addition to understanding this basic information, we will also learn more about him through long-term contact, such as:

• Are you an extrovert or an introvert?
• Do you have a catchphrase? What is it?
• What do you like to do in your free time, stay at home or go out and play?
• What are your food preferences? Do you like sweet or spicy food?
• Do you have many good friends? Are you popular?

…

And so on.

This information, which takes some time to obtain, is not basic information, but is closer to some privacy and preference information.

Today, as more and more users use our products, we will gradually improve our understanding of them.

In this process, we can use some means to label the user's habits, behaviors, and attributes. In the end, the user may become like this:

What’s the use of getting a bunch of labels like this?

Refined operations.

In essence, advertising recommendations, content distribution , and event marketing all require this set of things as a foundation.

Make good use of tags

In the entire portrait process, labels are actually a must-have.

When it comes to labels, many people may have a headache because many people think that how to type labels is a problem, but in fact, it is not that complicated.

The first step is classification.

Assuming that our operations are for C-end customers, we can simply divide them into several categories:

1. Human attribute tags

That is, using the information left by the user in the system, we can label his personal attributes in various ways, such as:

Male and female are labels in the gender classification; those born in the 90s and 80s are labels in the age classification; those born in the 1980s and 1990s are labels in the birthday classification…

2. Behavior attribute tags

That is, we can label the behavior attributes of users in the system, for example:

If there is consumption behavior, we can classify labels based on the time dimension and the single consumption amount or the cumulative consumption amount. For example, the rich are users whose cumulative consumption amount exceeds 100,000 yuan or the single consumption amount exceeds 10,000 yuan within 360 days.

3. Social attribute tags

That is, the various labels about the social attributes left by users in social networks and systems, such as:

A big V is a user who is followed by more than 50,000 people in the community; a heartthrob may be classified into this type because 80% of his followers are of the opposite sex.

…

You can also continue to type other types of labels.

To put it simply, the label itself is just a condition for classification. Even without a label, as long as you can clearly explain the relationship between behaviors and the superposition of required conditions, I think it is not a big problem. But experience tells me that if this is the case, it will restrict the complexity of the requirements.

Focus on use

Brother Liang has always believed that growth hacking is actually systematic operation.

The gap between growth hackers and the operations of most companies is actually the use of data, the speed of obtaining results through data verification after establishing hypotheses, and the ability to continuously verify data.

Because, whether it is user needs or user roles that we depict through Profile, in fact, it can be almost 100% considered to be lagging behind.

What we actually want to achieve with these lagged data is prediction, otherwise we won’t be able to get business opportunities.

Let's look at it this way. A user A bought 5 kilograms of apples on the platform today. When will this user have the need to buy apples next time? And how can we ensure that he will buy apples on our platform next time?

If you just think about it theoretically.

We basically have the following thinking model:

1. Aggregate all users who have bought fruit on the platform
2. Select users who have purchased more than N times in a row (assuming N=3)
3. Take the time of each purchase of these users and calculate the interval
4. Compare with user A, find the approximate value, and then make recommendations to A at this time interval to verify the effect.

Or change your thinking:

1. Through data analysis, let's calculate how many people are there in Family A.
2. Calculate how many apples there are in 5 kilograms.
3. Assuming that everyone eats one apple a day, how many days will it take to finish eating all the apples?
4. Change the assumptions to get different days to finish eating
5. Make recommendations to user A on different days to verify the effect.

You will find that predicting demand is much more difficult than judging demand.

In actual practice, I believe that if you simply recommend Apple, it may not be enough. To consider people’s decision-making process, you need to have a deeper understanding of human nature.

And this is another topic.

This article was compiled and published by @张亮 (Qinggua Media). Please indicate the author information and source when reprinting!

Product promotion services: APP promotion services Advertising