67 replies to this topic

[Adrian]

I'm having a read of Call to Action by Brian and Jeffrey Eisenberg at the moment, and in a section where they talk about 'uncovery', or working out who you are selling to, and what it is that you actually offer them.

They suggest building relatively detailed personas, offering caution about steteotyping, and going so far as to give these persona's full names, jobs, families etc... producing a description of at least 3 paragraphs.

Now, they specifically talk about creating detailed personas to avoid oversimplified stereotypes and superfiscial details that don't really help you. But then I start thinking, that because EVERYONE is different, EVERYONE fill up a continuos line of differences, with few that you can really 'pigeon hole'. So if you make 1 detailed persona, are you missing out on 10 other people who, though relatively similar, are different enough for you to need to change your approach?

Is there a danger that you could end up trying to cover all bases by creating 100 different detailed personas, some of them fairly similar to each other?
Or do you end up having maybe 5 personas, and trying to cover the gaps between them?

I can see the benefit of using personas, but on the couple of occasions I've been presented with them I've scoffed because they seemed to offer narrow views of the audience.

How do you go about thinking up useful user personas?

Edited by Adrian, 31 October 2007 - 08:44 AM.

[eKstreme]

When you develop personas, they should be at the extremes of the continuous line of differences. This way, you can clearly define the person you're targeting. Adding jobs and families is a useful exercise for consumer-focused products but not always necessary (e.g. selling an IDE to developers).

How to develop them? It's a bit of an art! I start by observing people around me if appropriate or trying to classify keywords lists into groups; e.g. keyword phrases containing the financial modifiers like "free" or "cheap" or those focused on quality like "best" or "top" or those at various stages in the buying cycle using words like "review" or "buy now" or "download".

Pierre

[Respree]

Maybe we can take about five steps back first.

What is a persona?

[Adrian]

Effectively it's a representation of someone who might visit your site. You create a fictional character, but one that hopefully represents some of the people who would visit your site, and then you use them to think about how they might interact with the site, and what you can do to improve the site for people like them.

In the book they use the American Heritage Dictionary definition of:

A voice or character representing the speaker in a literary work; The characters in a dramatic or literary work; The role that one assumes or displays in public or society; one's public image or personality, as distinguished from the inner self.

[cre8pc]

User personas originated with Cooper, in an attempt to design for real people.

Run a search on Kim Goodwin. She's a great resource and uses plain language. Here is an interview by UIE - Personas and Goal-Directed Design

There have been several variations of creating and implementing user personas. Obviously, the best ones come from some real data on users based on surveys with demographics, but this info is hard to obtain and never cheap.

Digging through usability studies is another way (how I do it), and looking for behaviorial studies, human factors, eye tracking...anything that offers insight into how people interact with web sites.

FutureNow, specifically the Eisenberg brothers and their team, have devised 4 personalities they use to guide design. My favorite is the "Methodical", which is what I use the most in my usability reviews. This person doesn't take any crap and has no patience. If you can sell to them or offer them an effortless experience, they return, make referrals, and convert.

One technique is the "Storytelling" technique. This is where you create a character. I compare this to acting in a play or acting class and you've just been asked to "Play the part of the multi-tasking divorced mother of 6 who has about 3 seconds to search for online courses." You describe her state of mind, emotional state, environment she's working in, type of computer, OS, browser, resolution, how many times the phone rang while she was searching with a baby on her knee and another kid is asking for math homework help..."

It can be a valuable tool in understanding that most people don't sit in a cube totally focused on their PC with their headphones on.

It's when you apply this character to a design that it gets insightful. Distractions in the environment make it difficult to stay on task and if the navigation or click paths are complicated, this person has far less time to deal with it.

Accessibility is another huge area. A user persona with ADHD is a potential user. So yes, there are limitless user personas that can be created.

This is why I like personalities approach by FutureNow. Global sites also need to consider how someone from another country may interact with their site. If no actual people are available or user testing is out of the budget, coming up with situations and characters helps flush out design issues.

I use user personas in my work as a teaching tool because often this is the first time a company has ever considered that people use their site differently than they do.

[Autocrat]

I'd go for 50-50 between detailed personas and generic user arhetypes etc.
As you suggested, getting to narrow in the definition leaves alot of others out of the picture... where as trying to be generically encompassing will jsut blanket paint everyone.

So, I doubt if you can get 100% reactive resposne, but you should, in short, get enough detail from your market to build a slightly less than generic idea of your visitors.
I personally don't see everytihng applying to everyone... but in general, there will always be Needs, Wants and Desires.
Meeting each of those groups will increase the liekly hood of sucessfuly convertions, and meeting several objectives within each of those groups is liekely to generate repeat custom and recommendations.


Just seems there are so many different ways to reah the same goal... and any times it's simple a slightly different incomming angle with a different label.
The upshot is, Know your Targets and your Markets - howyou label it or get to that is up to you.

[send2paul]

Adrian - hi

I knew we'd discussed this a few times before, so I dug around the archives and came up with a few conversations which might be of help, (if you haven't seen them already)...

1. Tool to help come up with great copy: Site Persona

2. User Personas - Intro to them - started by Kim! - which included this very useful link: An introduction to personas and how to create them (It's from March 2004, so some of the info might be dated).

It's all very intriguing this persona business... that's about all I can add to this debate

[iamlost]

The reasons for developing some 'awareness' of your traffic are (1) you design for them and not yourself, (2) focus down from the 'whole world' to the best conversion targets, and (3) build a structure to define your analytics data collection and use.

Most of us do not run focus groups, user testing, etc. but we do collect (or certainly should) web site visitor data (analytics). There is also an amazing amount of general (even free!) demographic data available, in print and online, if one is willing to research, extrapolate, and synthesise.

I view personas as various virtual visitor creations (alliteration!) representing 'who' visits a site/niche. My personas are real simple 'visitor demographic combinations': graphing statistical datapoint overlaps, i.e. broadband and dial-up; urban, suburban, rural, and unknown; geolocation age, sex, income, and social mores by percentage; niche interest levels by age, sex, income, location percentages; adaptive technologies used; etc.

Typically, I end up with eight to twelve obvious combinations (it tends to grow with time) that I label 'personas'. Crude but effective.

Then I run these personas through site scenarios. Basically running virtual behaviour drills noting why they come (from where, looking for what) and the impressions/problems they encounter (given their persona capabilities, disabilities, and biases).

The first site for which I built these virtual guests and drills seemed in constant design flux until I realised that I was putting too much 'me' into the scenarios. It was extremely difficult to separate what 'I, the designer' knew about the site from that of the persona.

If I was starting over I would have Kim and a few other specialists on speed dial - there is real value in competent third party testing. If budget is a concern (!) I recommend getting a friend or three to help develop the scenarios and to totally run the drills. Note: unless you go the stats route as I did I strongly recommend working with someone of the opposite sex to develop narrative personas.

I have done this now for all my sites and (1) it got easier and faster with each one; (2) subsequently fewer 'barriers' are uncovered as prior lessons learned are incorporated; (3) time on site, page views, conversion percentage, visitor retention, ROI, etc. increased dramatically as 'found' barriers were addressed; (4) personas are very very very useful in determining which Social Media sites attract 'your' visitors - and which do not.

[Adrian]

FutureNow, specifically the Eisenberg brothers and their team, have devised 4 personalities they use to guide design. My favorite is the "Methodical", which is what I use the most in my usability reviews. This person doesn't take any crap and has no patience. If you can sell to them or offer them an effortless experience, they return, make referrals, and convert.

Heh, yeah, they've introduced those 'types' as they've been talking about personas. I had a bit of fun trying to work out which I was. Probably some mixture of Methodical and Competitive I reckon
And they are goo at pointing out that even with these guides to types of people, most of mixtures of the 4, and when dealing with your site/product, someone who is usually Methodical could act Humanistic, because of what it is. So it goes a bitbeyond knowing who your audience are.

focus down from the 'whole world' to the best conversion targets

That's a good point. And maybe what's caused me to scoff a bit at the couple of attempts I saw at personas at my last job. I got the impression they created personas of the people they wanted to visit the site, not necessarily the people who actually did, and perhaps not the people who actually helped generate the revenue.
It is fair enough to focus primarily on those who are going to convert, and then add in the extras to take account of people who may not directly convert, but who act as connectors to people who might become customers.

Most of us do not run focus groups, user testing, etc. but we do collect (or certainly should) web site visitor data (analytics). There is also an amazing amount of general (even free!) demographic data available, in print and online, if one is willing to research, extrapolate, and synthesise.

I think this is the point I'm kind of missing. I've read about the personas and jumped straight to thinking about creating them, who they would be, what would they be like. But without any references to pull from or evidence to back up the behaviour of the personas that could be created.

Thanks guys, useful input all round

And yeah, I remembered it had been discussed a couple times before, Paul, but I looked back over the top 3 or 4 pages of the Usability forum and couldn't see the threads, so thought maybe it was time to bring them up again

As I read a few more pages into the book, I get the impression they are going to answer some of the questions I've got anyway

[saschaeh]

I get the impression they are going to answer some of the questions I've got anyway

Well i would certainly love to know what answers you unearth.

[victor363]

Persona's are a pretty abstract concept: for anyone having a hard time grasping the 'big picture' behind them - or just to see them in action: read this post (2 parter) by Brian Eisenberg. This is not in the book 'Call to action' btw (it should be though).

Now, they specifically talk about creating detailed personas to avoid oversimplified stereotypes and superfiscial details that don't really help you. But then I start thinking, that because EVERYONE is different, EVERYONE fill up a continuous line of differences, with few that you can really 'pigeon hole'. So if you make 1 detailed persona, are you missing out on 10 other people who, though relatively similar, are different enough for you to need to change your approach?

I usually take any knowledge imparted by the Eisenberg brothers pretty seriously, considering they are the best marketer's in the world and all. However, I have always felt that their goal here was to contribute to the science of marketing with some standardized and measurable procedure's. However, just like some people can multiply huge numbers in their head; I'm sure some marketer's don't need to think of persona's when designing their site's.

A lot has been written (online) by futurenow about how they create their persona's and make the characters 'more real'. I think they do this so that they can empathize with them better. Have any friend's that you feel like "you know them better than they know themselves"? If so, you can emphasize with them and predict their actions or responses, just like you want to emphasize with your persona's and predict their actions.
.......I wonder how this compares to the character creation process of novelists and screenwriters?

Best regards,

Victor


PS: I would love to see someone post their persona's and associated landing pages. This is a grey area for me, and I need as much practice with it as I can get ( also, it is a lot easier to learn from other people's websites since 1) there is no pressure and 2) you aren't suffering from 'inside the bottle' syndrome aka tunnel vision.

[Ron Carnell]

I could actually recommend several good books on creating user personas, but I suspect the best of the bunch would be this one.

Seriously. No, I mean, really. :-)

There are three sections to Dynamic Characters and the more pragmatic marketers can safely jump to the third, titled "Characters and Plot." This section, some hundred pages long, explores the never ending interaction between character development and plot, or put differently, people and what they do.

Asking what kind of individuals would buy your product isn't greatly different than asking what kind of character would spend every waking moment hunting a big white whale. It's a process of building characters to fit the plot, but it works equally well in reverse. Once you know what it would take to tell the story of Moby Dick, you also know what that character would NEVER be willing to do (we know, for example, that Ahab won't be running any marathons between ocean voyages). Plot is driven by character because what a person will do is determined by who that person is.

The Eisenberg's, as best I know (I have three of their books), never use the term narrative marketing, but I think they nonetheless advocate it, even if only indirectly. Personally, I think everything is a story; that's just the way the human brain is wired. And, yea, that includes building fictional characters that come to life through narrative.

[cre8pc]

Interesting suggestion! I like it because my guess is that the book supports not stereotyping characters.

That's the thing about user personas and a recognized danger.

Certain habits and behaviors can be categorized easier than coming up with 2000 potential individual users. (FutureNow leans on personality tests for example.)

When I create them, I often pick the unexpected visitor. For example, on auto parts sites, I use women rather than men because women, regardless of their car part knowledge or experience, may be more likely to be in a very different environment when searching for them than their male counterparts. Their brains work differently, which is one of the things FutureNow is exceptional at understanding and applying to their work.

User personas allow you to think outside the box and really stretch out and reach all kinds of people.

Another example is a site I reviewed targeted to young people but it sold products parents might want to buy for their kids. The site design, images, look and feel was so specifically targeted to kids that it was downright rude to adults - who have the credit cards. My user persona was a mom looking for something for her son, that he said he wanted from that site. Did they design it to include her?

Site designers aren't known for thinking like this, nor are stakeholders, which is why user persona work can be so valuable.

[victor363]

When I create them, I often pick the unexpected visitor. For example, on auto parts sites, I use women rather than men because women, regardless of their car part knowledge or experience, may be more likely to be in a very different environment when searching for them than their male counterparts. Their brains work differently, which is one of the things FutureNow is exceptional at understanding and applying to their work.

I like this philosophy as it probably ensures usability for the lowest common denominator. I imagine it makes the information less predictable for the target users as well - always a good thing.



Ron,

Thanks for the very informative post and for putting that book on my radar. It's convenient to have a writer here to answer our questions

To create a believablle character, what do your feelings about them have to be like? I've heard that the best way to understand a character or person, is to be able to love them - what do you think of this statement?

[Autocrat]

Okay... I think I'm going to come in from the far end of the spectrum on this one.

If there is a new client, with a new site.... how do you go about explaining to them the benefits of this type of practice... and more difficult still, how do you profile people that don't technically exist yet?

Profiling general only works on existing subjexts - so unless I'm mistaken, this sort of thing cannot be done until the site is somewhat established and has some form of usage.

Ideas?

[victor363]

If there is a new client, with a new site.... how do you go about explaining to them the benefits of this type of practice... and more difficult still, how do you profile people that don't technically exist yet?

Persona creation is a neccessary step to matching site copy to user's personality type's. Speaking 'in a person's language' was established as an effective sales technique well before the 90's. Though honestly, I think creating persona's will help them whether you can pull this off or not.

Profiling general only works on existing subjexts - so unless I'm mistaken, this sort of thing cannot be done until the site is somewhat established and has some form of usage.

Sounds good in theory; but personally speaking, I have a much harder time working on site marketing and sales if I'm working on a site that I've been involved with over a period of time. Tunnel vision is a real thing - and it can make the simplest tasks seem ten times harder.

[iamlost]

Profiling general only works on existing subjexts -

But the site content and purpose already point the way.

To take Kim's example: a site selling something that interests kids:
* you (should/could) have demographic data for the target children - who will do the persuading.
* for their parent(s) - who will actually do the purchasing (and subsequent recommendation).
* for businesses that also may need such products/services, i.e. daycares, pre-schools, etc. who may also do purchasing (and subsequent recommendation).
* for the initial geo-market area - which will more finely focus the social mores and disposable incomes, etc. of the target children and parents.
* etc.

As the actual traffic begins to provide additional data it needs to be incorporated. First to more highly target the already arriving traffic; second to expand (repeatedly) the most viable fringe(s) to grow the target base - much like long-tail keywording.

As victor363 mentioned some people do this mentally without effort, most of us do it a little automatically. There are two really good reasons to create formal personas/scenarios: (1) documentation - as reminder to yourself and as knowledge to someone new coming in cold; (2) it is usually easier to spot inconsistencies and missed opportunities on the written page. And the one especially great reason for doing it at all is that it really can significantly increase ROI.

...

It all comes back to how well you know your business and how well you prepare - business plan anyone? Yah I know, shaddup about the plan already, this is Web2.0, we doan do no freaking plan. :mr_rant:

[Autocrat]

...victor363.../...iamlost...
Good valid points... but I still cringe at the idea of smaller businesses attepting to tackle this sort of thing.
Most companies, in the UK at least, have no idea of such things (I'm assuming similar else where in te world?).

Moderate companies etc.... well yes, most of those get to be how they are due to such planning and forethought... but Home industries, Mom & Pop etc. (is that how you folk say it?) etc. won't have a clue, won't have the budget, and will ahe a damned hard time trying to convey their ideas over to me.
I'm a dab hand at business, markets and sales... I enjoyed those jobs almost as much as this line of work... yet I do not know the various markets different cliets are attempting to attract.

I general provide "basics"... such asage groups, profession levels and income types... phrasing to match the class/social etc... but I cannot do all that work - nor would I want to.


...iamlost...
Never ever stop with the plan/ning - I still reckon it's one of the best appraoches to anything (along with lists and doodles)... I still run my busiess from my plan, and I've never needed to shift from it (so flamming large it encompasses most thigns - kinds of cheating really .



Also, on a completely different topic - why do people bother with Web 2.0 ?
Before I got here, I used to laugh and form certain opinions on those that use such a term... (It's been around long enough, and only got names after plenty of time had passed!)... yet some of you folk use it - so now I cannot use it as a baramoter for peoples knowledge any more

[iamlost]

Most companies, in the UK at least, have no idea of such things (I'm assuming similar else where in te world?).
...
...but Home industries, Mom & Pop etc. (is that how you folk say it?) etc. won't have a clue, won't have the budget, and will ahe a damned hard time trying to convey their ideas over to me.

Yah most small businesses everywhere are as clueless as most small web developers. Funny that.

'Formal' persona/scenario implimentation is really for (1) those building for themselves who see the value and (2) clients with the budget who see the value.

That said, simply knowing the skillset exists allows a basic ad hoc persona/scenario implimentation, i.e. appeal to kids who persuade, convert parents who buy, etc. To paraphrase Kim (and the Shadow), 'the site is for kids, why would we consider the parents?' - the developer with the persona skillset knows.

Perhaps include levels of persona/scenario building at various price points and add another competitive edge and revenue stream...

I still run my busiess from my plan, and I've never needed to shift from it (so flamming large it encompasses most thigns...

BRAVO ZULU!!!
:thumbs:

why do people bother with Web 2.0 ?

For the same reason people bother with toolbar PageRank :popcorn:

[cre8pc]

I wanted to also point out that user personas are used in usability testing and from the inception of design. They can be surprisingly detailed. So detailed, in fact, that they may seem like someone you know. A professionally developed user persona reads like a story, with lots of details and insight into this "person" and how they think, what they like, want, need, and believe in.

Personas are focused on experience based design. Professional ones (hiring out for professional user personas is very expensive) help in supporting user goals critical to the success of a business. Understanding these personas aids in improving marketing efforts (like the site marketed to kids but not their parents, who would be willing to purchase there). Personas determine where to spend your budget and help prioritize spending. Do you improve the application or content, for example?

What are user goals? What are their Internet habits? How do they use your product or service?

The ethnographic research alone is one reason why they're so expensive. How many do you create? For these answers, I aim for Cooper, Kim Goodwin, and their extensive work.

For small/med business, just taking it all down to basics is an eye opening experience. When you get through to a site owner via even one example of a use persona, it can really open their eyes. They rarely view their sites the same way afterwards