From TheFutureOf (28 Aug 08): Response to Jim Novo's 12 Jul 08 9:40am comment

(sorry, folks. I don't have Jim's comment available)

I'm posting it here as I'm including some images and I need to post rather than comment in order for images to show up.)

Thanks, Jim. Good to be engaged.

Yes, I'm familiar with Recency. We use some variants of the standard concept in our blog research and tools. I'll agree that Recency is a pervasive human behavioral model. Do I think Recency is the best link between our worlds? That would take a few more discussions and I'm happy to learn if it is. Note that I'm qualifying my agreement because I think the devil might be in the details on this one (something I hinted at in my comment to your post.


As you point out, the Recency model is very simple and I definitely agree that one can do better. My understanding is that it's very useful as a general metric and its usefulness decreases as one digs down and explores why and how recency occurs. Some good papers on this are Customer value modelling: Synthesis and extension proposals (Journal of Targeting, Measurement and Analysis for Marketing, Volume 11, Number 2, 1 September 2002 , pp. 124-147(24)), How to develop new approaches to RFM segmentation (Journal of Targeting, Measurement and Analysis for Marketing, Volume 13, Number 1, 1 September 2004, pp. 50-60(11) and Metacognition and learning about primacy and recency effects in free recall: The utilization of intrinsic and extrinsic cues when making judgments of learning (Memory & Cognition, Volume 36, Number 2, March 2008, pp. 429-437(9)). The list is extremely extensive and I'm happy that a great deal of it either correlates, substantiates or validates NextStage's research. For example, I've talked at eMetrics and elsewhere about getting visitors to do things online simply by placing action items on pages in a way that makes volition near mandatory. Primacy and Recency Effects on Clicking Behavior (Journal of Computer-Mediated Communication, Volume 11, Number 2, January 2006, pp. 522-535(14)) is completely independent of NextStage research and does just that (I wrote about priming in my BizMediaScience blog along with related subjects).

For example, recency is highly age dependant with the strongest cues being in early adulthood (barring social impetus effects). This is great for people marketing to that demographic and the results shouldn't be thought as characteristic for other groups (see Temporal distribution of favourite books, movies, and records: Differential encoding and re-sampling (Memory, Volume 15, Number 7, October 2007, pp. 755-767(13)), Aging and contextual binding: Modeling recency and lag recency effects with the temporal context model (Psychonomic Bulletin & Review, Volume 13, Number 3, June 2006, pp. 439-445(7)) or Age-related differences in advertising: Recall and persuasion (Journal of Targeting, Measurement and Analysis for Marketing, Volume 13, Number 1, 1 September 2004 , pp. 7-20(14))). What I haven't seen in the literature so far is anyone using PCP methods to augment recency analysis (and I'm thrilled for someone to point this juncture out to me).


As far as a given model's simplicity being its greatest advocate for use…Have you seen From TheFutureOf (19 Aug 08): Response to Visitor Engagement Time for a reality check, specifically the part about not thinking that a simple metric is the best metric? In a lot of ways the quest for simple web metrics makes me think of the evolution of software itself. Originally not easy at all, then tools came along that made it easier and now it's a often a matter of point and click simplicity. But god forbid you attempt to get under the covers of the tool that's giving you that point and click simplicity! The post you offered at the end of your comment and the comments in that post indicate (to me) that simplicity wasn't a concern when developing those protocols (and I'm happy to be told otherwise).

I think this talks directly to your “…the pure simplicity of the model is tremendously appealing, especially when faced with the challenge of trying to get people to analyze anything at all.” The goal should be to provide simple to use metrics that demonstrate actionable results (and yes, part of that is understanding and developing metrics that aren't actionable at present or aren't financially successful (a nod to your honesty, Good Sir!)) because in the end (I believe) people don't care about what is being measured or how, they only want to to know how to get the effects they want.

I also recognize that I tend to resist Recency as a “primary screener” because I've seen the model fail more often than I've seen it succeed. This is probably due to NextStage making use of recency models to analyze how people process time (as noted in Responding to Christopher Berry's “A Vexing Problem, Part 4” Post, Part 2). People are “…are most strongly influenced by our most recent experiences…” so (in my world) recency is really indicating how much of a challenge a given population is having getting things into and out of memory in order to act upon them because people reference “recent” by markers their brains create in memory (you don't want me to include citations, do you? I already got flack once today from somebody looking over my shoulder at this and cracking up at the citations included thus far).

I learned that editors of an online I use to write for had an office pool based on how many links I'd include in my columns. It seems being able to document one's discovery path isn't as required as it use to be. Sigh.

This discussion interweaves with my Responding to Jim Novos 19 July 08 9:33am comment, I think, because one of our reports, Return Visitor Ratio, reports on how many visitors believe they'll be returning. The advanced version of the report includes a time calibration (similar to our Loyalty report) so perhaps I should share them in more detail.

Visitor Return Ratio, Loyalty, Credibility and Believability

NextStage Analytics Basic Visitor Return Ratio ReportThe figure on the right is one of our basic Visitor Return Ratio charts taken on 25 Aug 08. It indicates that about 78% of the people to this site will return (that's the black) and about 22% won't (that's the red). The little yellow dot on the left of the chart is an indication of how strongly the last visit is influencing the decision to return (basically not at all).

How come this success or failure of this visit isn't influencing the decision to return? Because the visitors are loyal and a single experience (unless extreme) isn't going to influence them that much (this hearkens to the definition of loyalty given in From TheFutureOf (19 Aug 08): Responding to Jim Novo’s 19 July 08 9:33am comment).

NextStage Analytics Loyalty ReportThe question then is “How loyal are visitors?” This is answered in the chart on the right. They are loyal and not outrageously so (the black bar is on the positive side of “0” and not a lot). The red dot on the “0” point indicates that although loyal, it won't take much for people to defect to a competitor. This is one way of recognizing when people are loyal due to something they believe or something they accept as credible.

Believability comes from a deeper place in the psyche than credibility. In a PersonalLifeMedia post to Moxie Insights SVP of Customer Insights, Diana Middleton I explained that believability and credibility work quite differently in human consciousness (and I'll include a link in the comments if the post hasn't gone live yet).

Credibility deals with facts and Believability deals with emotions and desires. Literally we're dealing with the differences between scientific thinking (Credibility) and anecdotal thinking (Believability). Again, this is something I wrote about in Responding to Christopher Berry's “A Vexing Problem, Part 4” Post, Part 1, My Easter Eggs Critiqued, Defining Engagement (Again? Oh, Lordy!) and Exploring the Holes in Flawed Logic and From TheFutureOf (16 Jul 08): Responses to Geertz, Papadakis and others, 5 Feb 08.

Credibility is easily measured; check references, check sources, interview/survey other individuals who participated or witnessed, etc. Litigation specialists and trial attorneys know the value of a credible witness (and “expert” witness) is (can you guess?) negligible. (This is something I mentioned during a dinner discussion arranged by Susan Bratton in Chicago and no, there is no link to it. Darn, huh?)

Loyalty due to belief is often demonstrated by that red dot on the Loyalty chart being way over to the right, well beyond the end of the black bar. What this particular chart is showing is that the majority of visitors find the information presented credible and not understandable (any guesses as to which site that chart is from?).

The true challenge comes when one wants to answer “When will a visitor return to a site?” People don't conceive of time as in hours or days or some such. At the DC eMetrics 07 I sat with two friends in the bar and put them through an exercise that demonstrated just how difficult it was for people to “think” of something much beyond 2-3 hours in the future or past, and that just thinking of doing something (actually thinking of doing it, not just saying “I'm going to do …”) much beyond a day in the future can exhaust most people's cognitive resources. This is because the majority of people surround themselves with a wall of “now-ness” and devote their attention to what's going on around them now (evolutionary installed neural wiring). This is why “planning for the future” is one of those wonderful things that so few people do truly well (I love Sir John Harvey-Jones' “Planning is an unnatural process; it is much more fun to do something. The nicest thing about not planning is that failure comes as a complete surprise, rather than being proceeded by a period of worry and depression.”)

The best approximation we use at present is that we have lots of understanding of what the majority of people recognize as “now” so we can guess when they'll return. I always emphasize to clients that it's a) a guess and b) doesn't take into account the elements of everyday life that prevent people from doing something they planned to do but wasn't something critical to their survival. What we're really reporting is when visitors believe they'll return (note, not “hope”, “think” or even “plan”. These imply conscious processes and while they're reportable they're not truly useful. I can “plan” (cognitive) on doing something but the conscious act (behavior) won't take place in reality until I non-consciously believe I have a reason (motivation. Hint to readers; motivations stem from the most primitive desires, the earliest wiring, and are usually the strongest drivers to action. This is why law enforcement always looks for “motivations” for a crime) to do it, hence the importance of the {C,B/e,M} matrix).

Guestimating when someone will do something is based on determining their intender status. NextStage's solution to this was developed 3-4 years ago and is based on Husserlian Conceptual Mechanics. You can find references to either in any of Priming the Conversion Pump with Color, The First Sale is the Next Page: an NSE Marketing Paper or Empathy and consciousness.


Intender Status

Intender Status Chart ExampleSo anyway, this is what a typical Intender Status report looks like. What it's showing is that the majority of people coming to this site believe they'll return the same day (red bar), within the next 24 hours (green bar) or so far down the road they really can't comprehend it (the white bar, labeled “> 6 months). Automotive manufacturers, governments, vacation properties and such, basically any organization with long decision cycles tend to be interested in the blue (1-3 months) and white bars, everybody else is interested in the red, green and yellow (1-3 days).

I appreciate that many people may be looking at these charts and wondering “That's useful how?” First, please remember that NextStage came into being to prove the technology in our patent worked hence many charts are based on being able to predict, cause and propogate behavior. We're changing our focus at present (scary, huh?). Second, yes, these charts do quickly lead to suggestions and action paths for improving marketing results. Just takes a little training (surprise!).

Tying this kind of data to a Recency metric would be interesting, I think. It would probably point to a way of being able to cause someone to return to a site at a precalculated point in time (and I'm already thinking of looking into certain sociology fields as I'm sure the research has been done there). Do you happen to have some recency models from small to large data sets you (or anybody) can share with me?

I think and am not sure that we're back to trans-temporal reafference, something I first entered into the discussion in a comment to you in a previous post.

It's funny (to me, anyways). You mention that “…you can improve the probabilities by adding other information — past products purchased, number of customer service calls, etc. — but the primary screener is still the Recency effect.” I suppose (and this isn't meant as a criticism, only as my mind pulling stuff together) that Recency as you've defined it as an excellent “big hammer” metric. i'll recognize and honor that, in fact.

My studies, though…I've seen and NextStage has agonizing amounts of data on how different types of “noise” — social networks, fair-exchange, … god, the list is quite long — affect people's behaviors.

I think and am not sure that this discussion leads to a simpler way to deal with what you call “dis-engagement”, recognizing it sooner (or at least through a variety of measurements. I'm a strong believer in triangulating for accuracy. If you look at some of my math you'd see that I often dodecalate and have been known to go much higher to improve accuracy and surety) and alerting stakeholders of it.

'Recency with Modifier', Pain-Pleasure, Negative and “0” Acceptance

I would agree with your “…this 'Recency with Modifier' situation as quite similar to the Pain-Pleasure, Accepting-Rejecting model…” et al. I disagree that 'Marketing will have no effect on people with “negative acceptance”.'

I think one of the places marketing must focus its attention is on consumers with negative acceptance because the strength of an individual's or group's negative acceptance indicates how much they're willing to act upon it (word of mouth, networking, etc).

Let me start by making sure our definitions are similar. Rejection (negative acceptance) is not the opposite of acceptance. The opposite of rejection and acceptance is indifference. This is similar to “What is the opposite of love?” The opposite of love is apathy. Apathy is also the opposite of hate. Both love and hate can be very strong emotions that direct a wide breadth of our neural resources towards individuals, groups, social systems, clan affiliations, so on and so forth. The opposite of a strong emotion is a lack of emotion. The height that one can love will always be equaled by the depth one can hate (see Where You Should Stick Your Ad and Why).

Similarly, the height that one can accept something will always be equaled by the depth one can reject it. Marketing will have its toughest job with people who are at the “0” point. Something needs to happen to shake them from their indifference. Current and historical research indicates this is where social marketing methods truly shine. Also, these methods are so culturally specific as to allow near surgical precision in getting messages out (here the citation list is agonizingly long. Let me know if you're interested and I'll post a bibliography of others' and NextStage's research on BizMediaScience. Some quick reads would be the first section of Reading Virtual Minds, my posts on PersonalLifeMedia or my Social Network or Word of Mouth BizMediaScience archives).

“…acceptance is not static…”

You write “…acceptance is not static, it decays over time. And that is why the timing is so important, you have you have to get in front of decaying acceptance (dis-engagement) and act before the customer slides into negative acceptance. Otherwise, it often becomes too expensive from a Marketing perspective to reverse the acceptance. There is probably psych literature on this idea of 'resistance to reversing acceptance' though I am not aware of it. When I've seen this effect, it often looks physical to me, as in 'a body in motion tends to stay in motion'. This is why I often refer to 'Customer Momentum', and it manifests in the parabolic nature of the graphs in the examples I provided.

Well stated and well said. The reasons for rejection/negative acceptance are fairly well understood, easily recognized, so on and so forth. It does take some effort to turn rejection/negative acceptance into acceptance because the neural mechanism is more like a switch than a slide (your “too expensive” statement). Moving from the “0” point to either side is a slide in neuropathy and this is why it's a different problem to solve (and also why social marketing works so well in getting people's attention but not so well at keeping people's attention). The best description I can offer for the switch from rejection to acceptance is that marketers need to overcome a cognitive inertia in the prospects' minds (your “Customer Momentum” and I'm so flattered you used a term so similar to my own, momentum and inertia). This cognitive inertia/customer momentum occurs because people tend to think they have an “idea” of why they accept or reject something and most of the time acceptance and rejection are based on emotional, or “belief” constructs (and this gets us back to my discussions with Chris Berry (isn't it wonderful how inter-related all these things are? I love it). Marketing is powerless to turn a rejection into an acceptance unless it addresses the emotional cause of the rejection first (yes, you are correct, the literature is rich on this subject. Let me know if you want a research bibliography and I'll make one available on BizMediaScience).

Obviously we're in strong agreement although our language might differ. The one flag on the play I'll offer is to your “…acceptance is not static, it decays over time.” I believe I understand what you mean. My rephrasing would be “action based on acceptance decays over time.” I offer that because once something is “accepted” it tends to stay there unless there's lots of effort to dislodge it.

For example (and parents, tell your children to leave the room during the next paragraph), most of us never gave up an acceptance of “Santa Claus”. How we demonstrate (act on) that acceptance has changed, perhaps, and rarely the acceptance of the concept. We went from opening presents to giving them, from sitting on the elf's knee to being the elf. The spectrums involved require us to accept “Santa Claus” throughout, merely to act upon that acceptance differently (okay, kids back in the room).

Yes, strong agreement that these effects occur everywhere. They're part of human nature/psyche/wiring. If they didn't occur I'd be concerned. It would be evidence of pod people (here I write of “Invasion of the Body Snatchers”, not people with iPods and such… and is there much of a difference sometimes?)

Yes, again strong agreement with your “…the more control a customer has over the interactions, the smoother (and thus more predictable) the response functions are.” This goes directly to my SNCR presentations about the history of technology being a demonstration of placing power in people's hands. Love your examples, by the way, and especially the “Friction” comments. That will bear some investigating.

I did read your Many thanks for directing me there. I posted some thoughts there, as well. I'm seeing much overlap in what we do although not quite in how we do it; complimentary, corollary, congruence, … My two cents, anyway.

One thought on “From TheFutureOf (28 Aug 08): Response to Jim Novo's 12 Jul 08 9:40am comment

Comments are closed.