Question:
What exactly is machine-learning as it relates to artificial intelligence?

Answer:
Machine learning, at the intersection of technology and behavioral research, is a field of computer science that gives computers the ability to learn without being explicitly programmed. AI is broader. Anything that can make a machine or software do something that would usually require a thinking person is artificial intelligence.

Question:
What is a Persona?

Answer:
A Persona is a machine learning agent trained to have interests that match a person or group of people. A Persona can represent a particular type of consumer, for example, or it can represent a cluster of consumers similar to what happens in segmentation research. Segmentation algorithms apply machine learning techniques to data to create clusters of respondent segments with similar interests. These clusters and their set of corresponding interests are called Personas.

Question:
What is an important difference between a human and a Persona in terms of the way each behaves?

Answer:
An important difference is the large volume of information a Persona reviews every day. While a human is exposed to a small subset of the Internet and reviews information sporadically, a Persona examines large amounts of content every day to determine whether or not it is interested. Essentially, a Persona speeds up the learning process compared to a human since more information is reviewed in a shorter period of time.

Question:
What is an important similarity between a human and a Persona in terms of the way each behaves?

Answer:
Just like humans, Personas’ current interests and reading material influence what they look at in the future. Consequently, their interests can change over time, as happens with humans. Sometimes, as is also the case with humans, Personas are relatively stable since there may be little variance in their reading. At other times, Personas get “excited” about new topics that are related to their interests. As this exciting topic starts to age, Personas, like humans, move on to other things.

Question:
How are the individual characteristics of a Persona determined?

Answer:
We begin by having discussions with you, our client, so we have a clear idea of your research goals. Once your research objectives are determined, we work with our AI partner, Tanjo, to create Personas through advanced clustering techniques.

Question:
How do Personas go out and learn once they’re created?

Answer:
Once Personas are created, they learn in real time, scanning the web and viewing thousands of pieces of content that are most interesting to them. Personas are affected by what they see, and that informs their interests going forward. Personas have interests that evolve based on what they are exposed to and the relative popularity of the material they have seen. The more popular and aligned a piece of content is with the Persona’s interests, the more likely that Persona is to be influenced by it and to change over time.

Question:
Do we designate all the sites where we want Personas to learn or do Personas operate on their own once they are given a foundation of sites?

Answer:
The Personas operate on their own once they have been created. although we do have the ability to direct them to read specific sources if our clients desire.

Question:
Why is it important to know what a Persona is viewing and how interested it is in the material?

Answer:
On its own, what a Persona is reading reveals current customer trends. More importantly, though, Personas incorporate the things they find interesting into their personalities, just as human beings do. This ensures that when we pose research questions to the Personas, we will get responses that reflect each Persona’s current view. This also ensures that we can conduct research with Personas at different periods of time to see if attitudes toward products are changing.

Question:
What is the “interest score” that is associated with Personas?

Answer:
The interest score of a message is a measure of conceptual alignment of the ideas and topics in the message and the set of interest areas of a Persona.  Personas can be built based on multiple data sets covering everything from demographics to purchase habits to survey responses and more. These data sets come together and form an array of “interest graphs” for each Persona (keep in mind, Personas can represent one individual or a group of individuals, depending on a client’s specifications). Once Personas are created, the machine learning process scrapes 30,000+ sources each day from the Internet and connects the most relevant content to the Persona, given the Persona’s particular interest graphs. An interest score for every piece of content is created by matching the content against the interest graph. The higher the interest score, the more closely it matches up against the interest graph.

Question:
Is there a way to understand what Personas read?

Answer:
Yes. We record everything that a Persona accesses, as long as the material has received an interest score of 75% or higher. A score below 75% means the material wasn’t popular enough with the Persona to impact its ongoing development. This behavior is similar to humans, who may stumble across an article that they think will interest them, only to toss it aside and forget about it when they realize that it wasn’t interesting after all. As Personas read, a list of the articles they are accessing along with each article’s interest score is produced, as is a word cloud that shows which topics are rising and dropping in importance. This information can be seen over a time scale of 24 hours, 7 days, 30 days, or the lifetime of the Persona.

Question:
If what a Persona reads can influence its reaction to research questions, is there a “filter” placed on this review process? In other words, does what it read 3 months ago have as much impact as what it read last night?

Answer:
Personas are like humans: their core values and personality traits will always have the greatest impact on their reaction to the research questions that are put in front of them, but information they read also can impact their decisions. The degree to which online material influences their current views is a combination of the degree of interest they have in the material and how recently the material was accessed. So articles with lower interest scores have less influence on the Persona’s current views than articles with high interest scores, just as articles that were accessed weeks or months ago have less influence than articles accessed yesterday.

Question:
Can Personas process sounds and images?

Answer:
No, not at this time, but it is a future goal. However, if images can be described in words, Personas can read the descriptions.

Question:
Can Personas access YouTube videos?

Answer:
Yes, but the videos will need transcriptions. If transcriptions are present, Personas can read them.

Question:
Can Personas access sites that require logins or subscriptions?

Answer:
Personas do not access sites that require logins or subscriptions. However, if a client needs particular sites with login requirement to be accessed, we can direct the Personas to do so as long as this does not violate the site’s Terms of Service.

Question:
Can Personas distinguish between advertisements and articles when they access a website?

Answer:
Yes. Personas are able to make this distinction and will not read advertisements.

Question:
Can you restrict Personas in terms of what you want them to learn?

Answer:
Yes, we can customize the journey of a Persona. For example, we can define the geographic regions they access to ensure information is reviewed from a global perspective.

Question:
As Personas mature, does what they read have less of an impact on the interest scores they produce?

Answer:
No, what Personas read can impact their interest scores regardless of how old they are. This is one of the benefits of Personas – they do not become static over time.

Question:
Do Personas learn at an even pace or, like humans, is there great variance in terms of the degree to which they acquire knowledge?

Answer:
Their learning is affected by what is going on in the wider web. For example, if a lot of activity is happening one day and less another within a Persona’s areas of interest, there will be variance in how the Persona is affected.

Question:
Do Personas access social media sites?

Answer:
No, these sites are not accessed because the posts are not reliably long enough.

Question:
Do Personas ever go to a site and then decide not to review something?

Answer:
No, if they access it, they review it. However, just like humans, sometimes they show more, or less, interest in what they’ve seen, so we only record sites where they have shown a statistically significant interest.

Question:
When a client’s concept is put in front of a Persona and the Persona gives the concept a “score,” is the score cumulative in nature (reflecting everything the Persona has experienced throughout its life)?

Answer:
No, the scores reflect whatever is going on with a Persona at that particular moment, similar to a human when survey questions are being answered at that moment-in-time. However, scores can be captured at specific time intervals to see how the Personas’ preferences are changing or staying consistent over time.

Question:
Can Personas go back to a prior point in time and respond to messaging or concepts being presented to them?

Answer:
Yes, you can “reset” a Persona to a particular point in time and anything that was learned after that point in time will not factor into a Persona’s evaluation of messaging or a concept. In other words, with a Persona, you can move along a time continuum, being able to compile data based on important events that have occurred. This is particularly important when thinking about product launches, for example.

Question:
When a Persona is exposed to a client message or idea, what is the “overall interest scores” that is produced?

Answer:
Just like humans, a Persona’s interest in a message or idea never occurs in isolation, but instead is influenced by all its experiences and preferences. In other words, a Persona’s interest in a message or idea occurs within the context of its “overall interests.” For example, when a Persona is exposed to a marketing campaign, the Persona takes into account all its interests before it decides if and how much it likes the campaign.

Question:
What company is behind the technology that produces Personas?

Answer:
PersonaPanels works with Tanjo, a dynamic leader in the field of artificial intelligence.

Question:
Has PersonaPanels established any partnerships at this time?

Answer:
Yes, PersonaPanels has been accepted in Nielsen’s Connect Partner Network and will be developing Personas for Nielsen clients using the vast Nielsen database. Nielsen’s Connect Partner Network is an elect group with inclusion based on innovation.

Question:
What is the relationship between Personas and humans?

Answer:
Personas representing humans mirror the behavioral patterns and expressed desires and preferences of actual humans – either individual humans or population segments, depending on a client’s needs. This technique results in an ever-learning AI representation of the desired audience. Once created, Personas are plugged into the Internet, where they read and view material appropriate to their interests. For segmentation research, the output from Personas is more accurately representative than the output from individual people because it relies on something closer to a census, so clients get a profile of behavior that cannot be obtained from small samples of respondents that may not be representative of their respective universes.

Question:
What are some of the advantages of conducting research through Personas as opposed to traditional research efforts?

Answer:
Let’s start with the important fact that Personas do not get fatigued so the number of questions they’re willing to answer and the length of time they devote to an exercise is not a factor. Next, traditional research uses relatively small sample sizes, whereas with Personas, we can move more toward a census. Third, there tends to be a variance in concentration among respondents. Personas do not have ebbs & flows in concentration levels. Fourth, we know that respondents can answer questions in one way and behave in another. This is never the case with Personas, which respond honestly and exhibit no tendency to try to “please the questioner.” Fifth, sometimes humans have issues that cause timeline delays in projects, resulting in missed deadlines. This will never happen with Personas. Sixth, using Personas for your research needs allows you to maximize privacy since sensitive topics are not being reviewed by humans. In traditional research, these sensitive topics can leak out in unintended ways, causing problems for the client company. Finally, let’s end with the fact that most research engagements represent a point in time when data was acquired through respondents. Although the hope is this data will have a nice long shelf-life, the truth is that, more often than not, the data ages quickly and its value to decision-making fades. Using Personas allows research tasks to be repeated as often as desired – with no expensive incentives being paid to respondents. Repeatable experiments result in better strategic decision-making.

Question:
What is the typical number of segments that can be created by using Personas?

Answer:
Traditional segmentation research usually features a fairly small number of segments (e.g., 5) but segmentation through Personas can go far beyond that number (e.g., 25), given that a census is approached as opposed to smaller sample sizes.

Question:
Can data from a prior traditional research effort be used to create Personas?

Answer:
Yes. PersonaPanels can take your data, create Personas and generate new data for you as often as you like.

Question:
Will the use of Personas provide opportunities to better “cut” data?

Answer:
Yes, because Personas allow you to move more toward a census and away from small sample sizes. This allows for more refinement of the data because the numbers will support it.

Question:
What is the process that allows my data to be turned into Personas?

Answer:
Your data will be turned into defined segments in consultation with the PersonaPanels staff and through the Tanjo machine-learning process. Once this is done, messaging and concepts can be put in front of Personas to get their reactions.

Question:
Can different data sets be combined and then used to construct Personas?

Answer:
Yes.

Question:
Can I create a customized Persona panel of consumers that I can tap into whenever I want to expose them to messaging or concepts?

Answer:
Yes.  The beauty of a customized Persona panel is that it is always available to react to information at any time and without the need for incentives. Plus, you are able to target narrow consumer segments if you choose. Narrow targeting using traditional research methods is time consuming & costly and sometimes not feasible.

Question:
Once Personas are created, how quickly do they begin to learn?

Answer:
Personas can begin to learn twenty-four hours after being created.

Question:
Once Personas are created, how quickly can messaging or concepts be evaluated by them?

Answer:
Once the Personas have been created and begin to acquire information, messaging and concepts can be put in front of them immediately.

Question:
Can the Persona’s interest in a concept be compared with a human’s interest in the same concept?

Answer:
Yes, the results can be compared as long as the Persona and the human represent a similar segment of the population. For example, we can put your ad in front of a consumer and in front of a Persona representing a similar consumer segment and create 5 quantiles (0-20, 21-40, 41-60, 61-80, 81-100) for their answers to fall into and then compare results.

Question:
Can advanced analytics using techniques such as discrete choice have an application with research that features Personas?

Answer:
Yes, Personas can react to the “stories” surrounding products and the attributes/levels that are featured for each of them, as is the case with conjoint or discrete choice. Because there is no respondent fatigue, a lengthy list of variables and their respective levels can be shown.

Question:
What types of traditional research lend themselves to the use of Personas?

Answer:
Traditional research encompasses areas such as awareness, trial, usage (ATU) research, positioning, concept evaluation, messaging, conjoint, discrete choice, segmentation, pricing, and customer satisfaction research. These all represent opportunities to have work done through Personas.

Question:
Are there certain types of research that would not be appropriate for Personas?

Answer:
Yes. One example would be message recall work with respondents who saw sales representatives. This would not be an appropriate application for Personas.