People Don't Change, Demographics Don't Work: Q& A With Valuegraphics Database

We are who we are, forming our basic value system early in life, according to David Allison, founder of The Valuegraphics Database. His company recently launched a seminal study of how our value systems are impacting our reactions to the pandemic -- and, in understanding them, find ways to leverage shared values from very disparate groups to move forward.

Charlene Weisler: Please give me a short overview of the methodology, sampling etc.

David Allison: The values we uncover for a target audience come from our benchmark database of 500,000 surveys and 446 metrics, which is accurate (+/- 3.5%/95% confidence) for 180 countries. It's the world's first database of what everyone cares about. For the pandemic recovery study in particular (as with all the data we pull from the benchmark dataset for any client or question) we identify a statistical representation of a particular population and ask them enough questions for us to extrapolate a profile from the benchmark study. 



Weisler: Can you trend this study? Will it be continued after the pandemic?

Allison: The study results will not need to be updated. We measure core human values, which are formed during our late childhood/early adolescence, and do not change. Our lives are ruled by what we value: every behavior and emotion we have is a manifestation of some value or set of values. So, understanding the shared values of a target audience (there are only 56 possible values) is akin to having the OS for that group of people. You can unlock incredible passion and power because you know what they care about most.

Weisler: What were the biggest takeaways?

Allison: There are so many. They break into three main categories of findings:

1. Rallying the troops: We were able to identify what aspects of life people want to see recover from the pandemic -- industry sectors, community amenities, and more.

We know what the people who are passionate about professional sports teams care about most, for example, so one could solidify that base by talking to them about what they want to hear. We can do the same for tourism, small shops, restaurants, art galleries, etc.

Why is this useful? If you are trying to drum up grassroots support for, let's say, the hotel industry...what messages will land best with your supporters, and how can you win others to your cause? As grants, loans, tax abatements and policies are being structured to drive economic growth, the sectors with the most supporters will have the most sway over the decision-makers.

2. Taking advantage of new consumer demands: We were able to identify the new pandemically triggered behaviors that people found they actually enjoyed, and want to keep doing. Each of these (there are four big ones and a few less important ones) represents a new market opportunity for brands to create products and services to satisfy. Further, we know what each segment of consumers values most, so it's easy to see what you'd need to say and do to engage and influence their behavior.

3. Motivating action to save the planet: We identified a segment of the population that was willing to keep up some pandemic behaviors that they found inconvenient as long as they knew those efforts would benefit the planet. In other words, we identified the most likely to take action to help with the environmental crises. And, of course, we can point to what they want to hear -- what they are listening for -- because we know what they value most of all.

Weisler: How did responses break out by gender, region of the country, age, ethnicity, etc?

Allison: This gets to the crux of Valuegraphics. Demographics are a legacy method of understanding groups of people. Our benchmark study proves conclusively that except for external characteristics, demographic cohorts do not resemble each other in any appreciable way. For example, let's look at age.

[We have charts] that show how much the people within any age category agree with each other about anything we have measured, including our 56 core human values that determine every behavior and emotion. We have 380 OTHER metrics too (for a total of 436)… The alignment on ANY of these is incredibly low. It's a bit of a cliche, but it's true: Nobody acts their age anymore. It's time we stopped using demographic stereotypes to look at the world around us.

It's a bit of a mind-expanding construct the first time anyone encounters Valuegraphics, as it challenges thousands of years of demographic profiling. Demographics as a sortation system are from the olden days….and abandoning the idea that demographics are destiny is the last great disruption required for us to truly move past the stereotypes that cause so much divisiveness and hurt in our world today. Especially now that we have a global database that proves how inefficient they are. Demographics push us apart. Valuegraphics unite us.

Weisler: How do you think these attitudes will play out a year from today? Two years from today?

Allison: They will be exactly the same. We have before/after studies in a few industries that prove this. Our values are forever.

17 comments about "People Don't Change, Demographics Don't Work: Q& A With Valuegraphics Database".
Check to receive email when comments are posted.
  1. David Scardino from TV & Film Content Development, December 23, 2020 at 10:50 a.m.

    I'm sorry, but doesn't our current political polarization give the lie to all of this? If our core values never change, how do you explain the difference in reactions between the American election of 2000, a truly close race that came down to about 500 votes in one state, and what we're currently going thru following an election that was nowhere near as close. And you're telling me "our values are forever..."? Please! (And before answering, you may want to check out video of armed protestors smashing the doors of the Oregon state legislature. Jus' sayin')

  2. Ronald Kurtz from American Affluence Research Center replied, December 23, 2020 at 1:29 p.m.

    David Scardino is right. This whole article sounds very naieve and idealistic. 

  3. David Scardino from TV & Film Content Development, December 23, 2020 at 1:33 p.m.

    Thank you.

  4. John Grono from GAP Research, December 23, 2020 at 4:45 p.m.


    I note that their website says "Your target audience is hardwired to chase what they care about. It’s what humans do."

    Also from the website ... "We have collected close to a half-million surveys from around the world, using 152 languages. We asked people what they value, want, need and expect in life."  Fair enough.

    It goes on to say ... "This first-ever, purpose-built dataset of contextualized values is a random stratified statistical representation of the population. The information it yields is incredibly accurate."

    One of the responses in this article was ... "This gets to the crux of Valuegraphics. Demographics are a legacy method of understanding groups of people. Our benchmark study proves conclusively that except for external characteristics, demographic cohorts do not resemble each other in any appreciable way. "

    So, if 'demographic cohorts do not resemble each other in any appreciable way', but their data base of half a million was based on a 'random stratified statistical representation of the population' and that "Your target audience is hardwired to chase what they care about", doesn't that infer that either (i) their premise is suspect, (ii) their sample collection basis was at odds to their premise, or (iii) the data generated from the data base was inheremtly biased towards respondents chasing what they care about?

  5. David Allison from The Valuegraphics Database replied, December 24, 2020 at 12:03 p.m.

    Hi David: Whatever behaviors and emotions people are exhibiting around the election, around race relations, (and, indeed, around all the issues that face humans in other parts of the world too) our values are the drivers behind those behaviors/emotions.

    The example I always use is the pandemic. If, pre-pandemic lockdowns, you were someone who placed an enormous value on family, you would have doubled-or-tripled your efforts to ensure your family was safe and secure. If however, family was not important to you, you wouldn't suddenly be worried about your family.

    We don't discard old values and grow new ones based on outside influences. We react to outside influences based on our values. 

  6. David Allison from The Valuegraphics Database replied, December 26, 2020 at 1:06 p.m.

    Ronald Kurtz, I'm not sure how to reply to your comment that the article sounds naive and idealistic. If you'd care to elaborate with an example or two? I'd be happy to try and explain whatever it is that makes you feel that way. 

  7. David Allison from The Valuegraphics Database replied, December 26, 2020 at 1:20 p.m.

    John Grono, I'm not entirely sure I'm following your logic here, but I think perhaps your point is that this demographically arranged dataset is being used to make a post-demographic point? If that's the case, this might help: 

    For any demographic cohort you care to mention (age, income, marital status, etc.) we can examine how often they agree on any of the 436 values, wants, needs and expectations we measured across 500K surveys. On average, the people within each cohort only agree on anything 10.5% of the time. Some specifics:

    • baby boomers agree 13% 

    • millennials agree 15% 

    • men agree 11%

    • women agree 10% 

    When you also consider that around the world the entire respondent pool - just by virtue of being human - agrees with each other roughly 8% of the time, the "lift" for using a demographic cohort to profile a target audience becomes negligible. 

    Our conclusion is that demographics are still useful as a way to understand what a group of people are, but they are not able to tell us who people are. Millennials do not all like avocado toast and boomers are not all terrible with technology. 

    I hope this helps


  8. David Scardino from TV & Film Content Development replied, December 26, 2020 at 1:27 p.m.

    David Allison, thank you for your response, but it still doesn't explain how, if values haven't changed or changed little, we get armed protestors attempting to break into the Oregon legislature or looking to kidnap the governor of Michigan in the aftermath of a not very close election as opposed to no violence after the much, much closer 2000 election.

  9. John Grono from GAP Research, December 26, 2020 at 5:59 p.m.

    Thank you DA.

    I'll try to simplify:
    - the article states "demographics don't work", yet demographics are the basis of the sampling which seems contradictory
    - agreement of opinions, values, beliefs etc. within demographic cohorts has always been low, thus demonstrating the necessity for demographics, for example favourite foods are very different between young and old (age), religion various by country (geography), fashion varies by gender
    - the article states " the world's first database of what everyone cares about", yet I have seen many databases of that ilk in many countries. Roy Morgan Research in Australia has used Values Segments for decades and is virtually ubiquitous in media and marketing in Australia.

    So, if your point is that targeting age/gender is a proxy then of course I agree, as that has been known for decades.

    But there are subtle differences between 'strategy' targets and 'implementation' targets'.

    For example a maerketers target may be 'intending F150 purchasers in the next six months'.   Good luck finding any plausible media metrics that you could trade on with large media owners in order to get the mass you want.   But you will probably find that 75% are males 25-50.   Yep, you can trade on that cohort.   So the strategist 'disects' the target (F150 being the 'Value') and gets the highest coverage level they can, which the buying implements.

    I hope that helps.

  10. David Allison from The Valuegraphics Database replied, December 28, 2020 at 8:52 a.m.

    David Scardino I hope I am putting these comments in the right place, I'm new to this site and it's a bit hard to figure out how to reply to a reply to a reply! Let me take a stab at your question using your specific examples, paraphrased below for brevity. 

    "Last election (close results) there was no violence. This election (wider margin) there was violence."  And your question is...."Doesn't this prove values have changed?" 

    I hope I've caught the essence there? 

    Here's my response. 

    Let's assume (because I don't know without launching a study) that there was a group of people who place a great deal of importance on Family, and on Meeting Basic Needs, just two of the 56 values that motivate people to do what they do.

    In the last election, the results may have caused some fear, and they felt these values were under threat. The manifestation of that fear/potential threat might have been letters to the editor and heated conversations at the coffee shop. I'm making this up, of course, but my point is that for whatever reason those values were under threat to some extent and it triggered some behaviors that seemed appropriate to those people.

    This time around, those same values felt absolutely on the verge of being destroyed...these people felt like there was a full-scale nuclear attack on those values, which triggered a far more violent response.

    But the values responsible in both cases were the same. 

    Also important to note that some of those who resorted to violence would have had different values that they were rising up to protect. Those angry crowds would not have all had exactly the same values in exactly the same order of importance. 

  11. David Allison from The Valuegraphics Database replied, December 28, 2020 at 10:03 a.m.

    John Grono, let me start with the same disclaimer as my last post: I am new to this site and not sure if I am responding in the right place here, but I hope this gets to you.

    Trying to reply to your comments in some semblance of order below. This may take a one-on-one conversation if we get much more granular, which I am happy to do if you like. Just let me know. Here we go: 

    Demographics remain an excellent tool to fence off people/ segment populations with similar outward-facing characteristics. It's a useful system to describe WHAT people are. But a lousy system for understanding WHO those people are. If you try to use demographics to describe WHO people are you are stereotyping, and you end up with sweeping false assertions that millennials love avocado toast, men like sports, and rich people like luxury hotels. 

    Example: I've done work with radio stations, where units of advertising are still sold by demographic segments. But when you talk to the programming people they shrug and ask "if we are supposed to be a station that appeals to 18-24-year-old women, what am I supposed to do with that? They don't all like the same music or want the same content." And that's the point. Demographics are useful to describe a group of people, but not to understand what makes them tick.

    For those radio stations, we determine what the particular 18-24-year-old women that they are attracting actually value, a.k.a. what they care about, and that helps the programming people figure out what to serve up to solidify the base. 

    The welcome side-benefit of this post-demographic (or perhaps supra-demographic is a better term?) profiling is that those stations using values-based programming end up attracting a lot of other people outside the demographic they thought they were focused on, because values are not correlated to age or gender or any other demographic label. No demographic owns a value more than any other. 

    ,,,continued in another post....

  12. David Allison from The Valuegraphics Database, December 28, 2020 at 10:11 a.m.

    PART 2 of 3

    You go on to mention demographic variances in food preference, religious affiliation and fashion. Couldn't agree more. Values are the root of those behaviors and decisions and are largely unaffected by demographics, while the manifestations of those values will vary from one demographic to another.

    Example: I'm very focused (my partner would say obsessed) with my food intake because of the importance I place on Health & Well-Being and Friendship, just two of the 56 values that might motivate my behavior. For me, that means I track my food intake with an app on my phone, I'm fairly consistent at meeting my protein intake goals every day, and I curtail sugars in any form. The value of Friendship figures into it because two of my friends also use this app, and it a kind of bonding thing between us, as we can see each others' daily inputs. We even place bets on goals from time to time.

    For someone else, the manifestation of Health & Well-Being and Friendship might be NOT drinking a bottle of wine every evening during and after dinner and going for a walk with a friend once a day. Both of these behaviors that originate in Health & Well-Being may very well be trackable by age or other demographic labels.

    What our dataset proves, however, is that the underlying values are NOT more prevalent in one demographic cohort vs. another. We are interested in and can isolate WHY a group of people do things, not what they do.

    For a marketer of a food-tracking app like the one I use, knowing that their target audience is driven by Health & Well-Being and Friendship opens up far more opportunities than knowing their target audience demographics or even current behaviors. They might stop promoting the tech-story so heavily (features/functions) and instead prioritize a positioning based on the Health & Well-Being benefits of tracking consumption patterns of all kinds. And they could come out with upgrades that expand the activities and intakes that can be tracked. Not the best example in the world, but I hope it helps make my point.

    I'm unfamiliar with the Australian Database you mention, but it sounds fascinating. And while "the first database of what everyone cares about" is a marketing phrase that necessarily leaves out some qualifying details, it's more useful for most people than the full version, which I do believe describes the first database of it's kind:

    The Valuegraphics Database is the first random stratified statistically representative sample of the population of nine regions of the world encompassing 180 countries built from 500,000 surveys in 152 languages and measuring the importance of 56 core human values plus 380 correlated wants, needs and expectations with a margin of error of +/- 3.5% and a confidence level of 95%.

    ...continued in next post.....

  13. David Allison from The Valuegraphics Database replied, December 28, 2020 at 10:12 a.m.

    Part 3 of 3

    Lastly, your closing example talks about F150's which I assume to be referencing trucks? As your example unfolds you mention that for the target group you've described, "F150's are the value" makes me wonder if we are on different pages because of nomenclature?

    "Value" is one of those words that means so many things to so many people. Marketers talk about brand values and product values and the ever-popular "value proposition."

    The values measured in the Valuegraphics Database are the core human values that motivate all behaviors and emotions: the overarching principles by which we run our lives. An F150 is not a value like that. There's a chart at the link below that shows all 56 values for the population of the world, in rank order.

    And again, I'm happy to discuss further if you like? Just let me know and we can find a time to connect.

    Happy New Year!


    Link to the chart of the 56 core human values rank-ordered for the world:

  14. Ed Papazian from Media Dynamics Inc, December 28, 2020 at 10:48 a.m.

    I have read the various comments on this thread with interest and have a few of my own. While it is absolutely true that age/sex is a very poor deliniator of anything it is not true that because that's all that Nielsen is asked to report ratings on that we are stuck with such primitive yardsticks. In fact, it is not true that media buys which use age/sex as their "currency" are actually "targeting" such conglomerate groupings. In the case of almost all TV and radio time buys these are merely broad audience "impressions" defining metrics---GRPs, usually---that allow the seller and buyer to agree on audience tonnage guarantees---nothing more.

    The plain fact is that most ad campaigns are based on product and brand positionning strategies which appeal to certain mindsets or consumer needs and desires---and these often transcend demos, even when the latter are better defined---like by cells such as age within income or edication. There are correltiuons, of course, so it's not accurate to say that any type of demographic analysis is wrong. But it's really the mindset that is what determines the outcome---and the ability of the advertiser to woo said mindset.

    John, I have employed studies such as are conducted by MRI as a very effective tool to establish the mindset of media audiences and how these can relate to ads where the primary appeal of the message is to people who are very health or style, or penny pinching, or whatever conscious and various combinations of same. This is primarily a media planning not a buying function---when the advertiser has the sense to insist of demanding that media planning and buying are truly integrated. The planners simply tell the buyers what genres to focus on---subect to CPMs---to generate the best fit---and the buyers use Nielsen's age/sex ratings to get their audience guarantee protection.

  15. John Grono from GAP Research, December 28, 2020 at 5:26 p.m.

    PART 1 of 2

    Thanks DA. I think the open discussion is valuable.

    Basically I agree with the premise. I chose F150 because everyone I know who has one is a welded-on fan. Yes, a very narrow definition, sorry. 'Apple Acolytes' may have been a better example.

    The big question (after using such data bases for many, many years) is, do consumers know how to enunciate their 'Values' as the majority don't think along these lines, unlike us researchers. We do our best to ask neutral questions, which more often than not return non-discriminatory responses. For example, 'Honesty' is a strong Value. And yep, that box gets ticked by most because everyone aspires to that. This has led to decades of an untold number of questionnaires, and I welcome your addition.

    And I 100% agree that age/gender/geography is a poor predictor of brand values, affiliation or intent to purchase etc.

    Most advertisers I have worked with describe their target with an assortment of adjectives and adverbs. For example, outgoing fashion-conscious well-to-do singles that are socially active. (And yep, I made that up, but I've seen many 'targets' like that). In the main, they use (rightly or wrongly) behavioural attributes rather than values to describe their consumers. Why? Because behaviour can change quicker than values (which also can and do change, but slowly), and their job is to increase brand sales in the short-term.

    In AU using data bases such as Morgan (n=55,000 p.a. mainly F2F) the strategists in media agencies forensically interrogate the data base and build a population profile primarily based on such behavioural responses (there are also 'belief' and ‘intender’ questions).

  16. John Grono from GAP Research, December 28, 2020 at 5:27 p.m.

    PART 2 of 2

    However, the advertisers' budget is finite.   The cost of reaching the outliers is astronomical.   Think of it as the left-hand and right-hand tails of the Normal Distribution.   The 'sweet-spot' is in the middle - the peak of the curve.

    We do a further analysis to produce coverage for various cohorts that make up the target 'universe'.   If you get 70% coverage you are doing well.   A typical cohort may be females 25-39, employed, high socio-economic, university educated and single (yep, made that up too, but it exists).

    So age/gender/geography can be a good predictor of such behaviours that the advertiser wants.

    Unfortunately that cohort is not part of the lingua franca, but the broader F25-39 is, but it blunts campaign efficacy.

    Fortunately, our Morgan data base has masses of media usage data.   So we 'filter' the media we select to invest in.   For example, it maybe be restricted to shopping mall OOH, some TV, premium magazines, and online.   Within each medium the individual locations, programmes, mastheads and sites are identified and THAT is what we trade on under the very broad umbrella of "F25-39".

    And no, I don't work for and never have worked for Morgan.   It's just that I've used it many, many times, and that in AU it is the most subscribed data base in that field.

  17. Ed Papazian from Media Dynamics Inc, December 28, 2020 at 7:40 p.m.

    As we all probably realize, there is a major disconnect between media buying---mainly for TV----and the kinds of mindset targeting that are the core premise of large numbers of branding ads. The disconnect is because the planners---who should be aware of the  targeting appeals to be used by their clients' ad messages---this is not always so, by the way----are largely shut out of the buying process. The latter at far too many companies is evaluated on a cost per set of eyeballs basis and, not surprisingly, the sellers and buyers are quite satisfied with this. The sellers get to move large numbers of GRPs attracted by mediocre programming---which they are otherwise stuck with---and the buyers don't have to get into finely nuanced demos---let alone ---to them---totally abstract issues such as the potential receptivity of each show's audience to the particular ad message the client intends to convey.

    As John notes, certain surveys that are available to planners and through them to brand managers exist for identifying with a fair amount of precision what shows or program genres attract above par percentages of particular mindset groupings---I mentioned MRI which is a fine source for such information. And, in the case of magazines, such data is often heeded. But not so with TV, which is the main event as far as most branding campaigns are concerned. Sadly, many CMOs and their brand managers operate on the assumption that a broadly defined 60% monthly reach ---which probably translates into a total year 85% reach--- is acceptable as this more or less guarantees that those who are most receptive to their message will be exposed often enough for it to have an effect. Beyond that, it's up to the commercials to seek out their audience is the prevailing thinking.

    My point is that we can debate this on a theoretical basis all we desire, but are the real decision makers---the brand managers and CMOs ----even willing to listen---let alone do something about it? How do we motivate them to participate in re-evaluating the over reliance on simplistic demographics by their media people?

Next story loading loading..