Re: [projectvrm] Personalisation WorksŠ up to the Uncanny Valley

[Katherine Kern < >]

Graham and T.Rob – the “uncanny valley effect” is brilliant.

But the most valuable number to me is the 4%.

Back in the day, “mass marketing” success formula for a new product was "80% awareness X 30% Trial X 70% retention = 16.8% adoption.” The variables are mass media (awareness), availability (trial), product performance (retention).  

If GfK’s “Brand attachment”(4%) is similar to “adoption”, what is the formula for success for a “personalisation” advertising campaign?  What if it looks like this: 10% Awareness X 60% Trial X 70% retention = 4.2% adoption.  I’m hypothesizing that “personalisation" advertising is so much more contextually relevant and well timed that trial is double a mass marketing campaign.  

As Graham points out it just may not be cost effective to invest in a level E “personalisation” tactic to improve from 4 to 4.8%.  Especially when there’s the risk of the valley effect if the “personalisation” is not just right.

How about reversing strategy.  Instead of a higher level of “personalisation”, target a wider audience (for example, a community with relatively common interests to the personalisation data) to increase the awareness level.  For example, increasing awareness from 10 to 15%, reducing trial to 50%, but maintaining the 70% retention rates yields 5.2% adoption (higher than the highest level at the E “personalisation” level, without risking the valley by getting too personal).  

How does this relate to VRM – maybe the entry level of a VRM tool invites individuals to share information that identifies them anonymously, relative to a community or a sub-segment within it, which Vendors can target with a message personalized to that context rather than the individual.  

K-

From: Graham Hill < "> >
Date: Friday, April 11, 2014 at 4:36 AM
To: Datar Sahi < "> >
Cc: ProjectVRM list < "> >
Subject: Re: [projectvrm] Personalisation Works… up to the Uncanny Valley

Hi Datar

I am with you 100% on this. 

There is no doubt that marketing personalisation works for most customers for the reasons you outline. From my vantage point of a project to optimise a retail insurer's marketing media mix, I can see that increasing personalisation increases the number of conversions, increases the number of quotes and increases the number of sales. Music to a marketer's ears. But that doesn't mean it is a get out of jail free card for marketers. There are two challenges that limit its ultimate effectiveness: over-personalisation and diminishing returns.

As T.Rob wrote about in an earlier post in the 'uncanny valley' (https://ioptconsulting.com/escaping-advertisings-uncanny-valley/) theoretical work in robotics applied to personalisation suggest that increasing marketing personalisation increases its effectiveness up to a point where it becomes so disconcertingly creepy that it rapidly declines in effectiveness. This is the uncanny valley. This was just a theory until recently when UK market researcher GfK carried out some research on the subject. As Colin Strong described in a presentation at the recent Personal Information Ecosystems event organised by Ctrl Shift in London (which I was fortunate enough to be invited to attend), as personalisation increased so did brand attachment (a proxy for marketing effectiveness) up to a point where the uncanny valley effect kicks in and brand attachment rapidly declines.

If the uncanny valley exists, and GfK's research tentatively suggests it does, the big question is how much personalisation has the maximum effectiveness. And related to that, how should personalised marketing be presented so that it's effect is maximised. My experience working with marketers suggest that we are still a long way from maximising the effectiveness of marketing personalisation.

Even if marketers knew what the optimum level of marketing personalisation was, that doesn't mean it is economically viable. As marketers seek to gather, analyse and generate usable insights from ever more data the associated costs increase at a much faster rate. Going from a level C of personalisation (on the graphic) to a level D might only require a modicum of customer data, analysis and insight and be cost effective, but going from a level D to a level E might require so much additional data, analysis and insight that it is no longer economically viable. Marketers are not yet widespread users of big data so much of these empirical calculations have yet to be confronted.

All in all, marketing personalisation works, of that there can be no doubt. The known unknowns are how much personalisation is optimum from both a marketing effectiveness and an economic viability perspective. Answers on a postcard to the usual address.

Best regards from Cologne, Graham

PS. Personalisation is just a proxy for inferred customer intent. But that is a topic for another long email.

On 9 Apr 2014, at 23:15, Datar Sahi < "> > wrote:

Hi Don,

Personalized ads make impressions (audiences) more valuable.

Example1: CNN.com has an ad on their homepage for home mortgage loan. Only .001% of folks that day are interested in a home loan.
- Data aggregators inform CNN based on behaviors, credit scores etc that they should only target 1000 visitors (impressions) to that site for that day. This frees up all other impressions (visitors) so they can receive different ads personalized to their data
Example2: Mazda knows you are in-market for a mid-size sedan. They know you are graduate degree educated and have high HHI and visit engineering blogs/content. They also know you are an environmentally friendly person.
- They serve you an ad that highlights the gas mileage
- They serve me an ad that highlights horsepower and cool factor based on my data
- They serve Doc an ad that highlights easy financing and local dealer (value and ease of purchase)
Each of the sub bullets above is a personalized ad for the same exact car and the advertising community exists to show value in the above.

The biggest cliche in the advertising: Show the right ad to the right person at the right time = Personalization.

Incidentally, all of the above can be done at scale and is already happening.

Datar



On Apr 9, 2014, at 10:18 AM, Don Marti < "> > wrote:

I'd split the categories into three, not two --
normally I'm a lumper not a splitter, but I think
it's important here.

Category      Offline example              Online example
---------------------------------------------------------
Search        Yellow Pages                 Google AdWords

Signaling     magazine ad, TV spot         Nobody (!)

Direct        direct mail, telemarketing   Everybody (!)

(Personalizing ads makes them less valuable, so
predicting that every ad will be personalized is the
same as predicting a collapse of the ad business.)

Don


begin Datar Sahi quotation of Wed, Apr 09, 2014 at 09:00:43AM -0700:

2 categories of ads here:

1 mass reach upper funnel: more content and awareness driven. No call to action and just reinforcing a brand.
2 personalized, refined targeting and stronger call to action

Marketers (businesses) need both. Without awareness, you can't achieve efficient actions.

There are very few mass reach plays today all of the fragmentation (consumption habits). Every ad in the future will be personalized.

Datar

On Apr 9, 2014, at 8:45 AM, Don Marti < "> > wrote:

Sure, but the Super Bowl as an ad medium was always
the most regulated, high-end, and "curated".  Yes,
the budgets and production values have gone up and up,
but there was never a time when Super Bowl ads were
as low-class as say, email spam, or the Facebook ads
that older women seem to get:
http://zgp.org/~dmarti/business/facebook-ads/

"Times Square outdoor advertising" is a category that
has gone from skeevy to high-class, but it seems like
a rare exception (that wouldn't have been possible
without the economic and legal changes in the whole
neighborhood, even the whole city).

Don

begin Doc Searls quotation of Wed, Apr 09, 2014 at 10:57:04AM -0400:

There are many distinctions to be made in the advertising business.

As for Times Square... I dunno. It might be a cultural attainment of some kind, but it's also an exceptions (bold and flashy electronic billboards) that don't generalize well to the whole category. Another is Super Bowl ads. People go to both so they can witness the best exemples of a category they otherwise ignore or avoid.

Doc

On Apr 9, 2014, at 10:46 AM, Jim Bursch < "> > wrote:

Some might argue that billboard advertising has attained a higher value as it has evolved with technology (Times Square comes to mind as another cultural attainment of advertising). And I suspect that billboards are the closest thing to a pure "ad medium" -- as opposed to some other medium to which ads are added. A distinction needs to be made between an "ad medium" and an "advertising-supported medium."
Jim Bursch
310-869-5340

">

@jimbursch

On 4/9/2014 7:28 AM, Doc Searls wrote:
I agree, though Vogue may be in a tie with other fashion/specialty magazines. The point is that the advertising itself is a kind of editorial, and adds value to the whole thing. Brand advertising at its best does that.

Still don't have an answer for Don's question, though.

Doc

On Apr 9, 2014, at 10:22 AM, Jim Bursch
< "> >
wrote:

I think of Vogue magazine as the highest cultural attainment of advertising.
Jim Bursch
310-869-5340


">
http://mymindshare.com


@jimbursch

On 4/8/2014 11:44 AM, Don Marti wrote:

begin Doc Searls quotation of Tue, Apr 08, 2014 at 11:36:20AM -0400:

A fun find in the comments:



<http://crappy-mobile-ads.tumblr.com>

Yes, that's great.  (My favorite is the clippable
coupon. Yes, just a second while I take a pair of
bolt cutters to my phone...)

Advertising history question: has there
ever been an ad medium that has gone from
crappy/spammy/disreputable to higher value?  Or do
ad media get burned through never to recover?

-----
No virus found in this message.
Checked by AVG -
www.avg.com

Version: 2014.0.4355 / Virus Database: 3882/7320 - Release Date: 04/09/14

--
Don Marti                    
http://zgp.org/~dmarti/
">

--
Don Marti                    
http://zgp.org/~dmarti/
">

--

Dr. Graham Hill

">

UK +44 7564 122 633

DE +49 170 487 6192

http://twitter.com/GrahamHill

http://www.linkedin.com/in/grahamhil

Partner

Optima Partners

http://www.optimapartners.co.uk

Senior Associate

Nyras Capital

http://www.nyras.co.uk