Join 3,561 readers in helping fund MetaFilter (Hide)


How effective is personalization of content?
March 25, 2014 4:41 PM   Subscribe

I work in advertising developing large corporate sites and the question I'm asking is very hard to Google as it is muddled in advertising speak. A big selling point in many of the large CMS systems that I work with are personalization: the idea that content can be delivered to users that meet certain criteria. I will go more into depth in what I mean inside, but the gist of my question is: is there any research that shows personalization is effective in creating "brand awareness," sales/conversions or anything else that can be tracked with a metric? I spend all day implementing these and just don't see the advantage.

For large ad networks I see the benefit. If I am on Google and Google knows I'm a ski enthusiast in a nice demographic, there's a pretty obvious point to targeted marketing. If I'm a brand that has no ecommerce function then personalization, while not worthless, begins to lose its appeal. Sure if I'm a large supplier of pillow stuffing and I find out my visitor happens to be a small business pillow stuffing buyer I can show them more small business pillow offerings in an internal side ad, but that seems pretty low. More realistically, if I'm a makeup manufacturer that does not sell directly online I might have a lot of visitors from a lot of demographics. I can surely target teens better than older adults, but how does that track if you're not linking personalized ads to direct purchases? You can just correlate aggregates. Furthermore, at what point can you say I have these many visitors, personalization makes sense at a certain level. Or A/B testing within a personalized demographic makes sense.

I guess I have a somewhat statistics/math background, or at least a background grounded in science, and while I know personalization makes sales of big $500k platforms easier (which is probably why they are such a big deal), I want to believe that there's something grounded in reality and a methodology for tracking these things.

This became a bit rambling, but all day I hear about personas, personalization, visitor profiles and I wonder if any of it is grounded in reality or we're just shilling. There's got to be some actual marketing research on this.
posted by geoff. to Computers & Internet (4 answers total) 14 users marked this as a favorite
 
the question I'm asking is very hard to Google as it is muddled in advertising speak.

I spend all day implementing these and just don't see the advantage.

I wonder if any of it is grounded in reality or we're just shilling.


I realize you're not talking about social media (necessarily) here, but this book argues that, as you say, marketers and advertisers have way overblown the value of personal profiling for businesses, mostly so they can sell *their* books and seminars on how to game the system. Definitely not a research-heavy book, but might be worth a look.
posted by Rykey at 5:50 PM on March 25


Advertising isn't (usually) focused on creating directly correlated purchases. It's about affecting brand metrics: brand awareness (do i know it exists), brand affinity (do i like it), brand preference (do i like it more than competitor brands), brand relevancy (is it a brand that matters to me), brand comprehension (do i know what that brand stands for). Personalization is mostly useful in that it puts the brand in front of people who are more likely to find the brand relevant, and are therefore more likely to pay attention to it, and are therefore more likely to have positive brand impressions. Also, personalization saves advertisers a lot of money by ensuring that you're not, for example, paying for showing your tampon ad to a 15 year old boy.

You're not seeing correlations because you're not measuring those things, but you can bet that your clients are, as are the marketing science folks at your agency.

(I'm a digital strategist, and i produce all those personas etc etc etc, and metrics from my projects seems to indicate it is not crap.)
posted by Kololo at 6:10 PM on March 25 [4 favorites]


While I don't have any links research off the top of my head, this is a very interesting subject, and I do tend to agree that there really ought to be a business case (and it is possible to investigate the business case) before implementing a relatively costly enterprise-level CMS system.

My background is local marketing and software marketing (B2B and B2C). In every vertical, customer personas are indeed very important in many ways (helping improve a software product, or helping understand why a guest is choosing a $3000/night luxury resort).

Peronas help understand user / purchaser intent, which helps optimize the "conversion funnel" on a website.

However, the businesses I work with are SME's and not Enterprise-level, so the number of users/customers will be relatively smaller.

So, it's quite easy to implement roll-your-own data collection (surveys, email and phone interviews) to find out more about buyer personas, and if needed, segment them.

However, with SME's, often the goal is to find the most relevant segment, and optimize for them first (as opposed to diluting efforts by optimizing for a number of personas).

But in terms of the cutoff for ROI for an actual platform, I would like to know too!
posted by KokuRyu at 6:12 PM on March 25


It's actually really unclear what you're asking.
I work in advertising developing large corporate sites and the question I'm asking is very hard to Google as it is muddled in advertising speak. A big selling point in many of the large CMS systems that I work with are personalization
...what kind of sites? And what kind of personalization? If you can say more about what you're thinking of, we might be able to answer better. It also seems like you're mixing up a bunch of concepts, i.e., targeting (by demographic) vs. personalization (based on signals about what the user might like, like Netflix.)

For example, with your pillow sales site, I don't really understand what you mean by showing an ad for small pillow sales companies, presumably you wouldn't show an ad for other companies. Also, you describe the site as internal, usually internal sites wouldn't have ads, as they serve a captive audience.

Regarding your makeup sales example, that seems like a pretty clear case for targeting. I.e., if you know your existing customers are women in their 30s, it makes sense to advertise toward that demographic, because it's generally easier to get people to do a little more of something they already do than it is to get totally new people to do something totally new. On the other hand, if you find you're no longer seeing gains targeting women in their 30s, then it makes sense to start targeting a new demographic. These campaigns are different though, i.e., if they already know about you, you're just reminding them, whereas if they've never heard of you, you're trying to just lay the ground work.

As far as measurement goes, you're right that it's tricky. One option is to advertise to a certain geographic area, and then look before and after the campaign, or to measure sales at stores in that campaign to see if there's any change. You can then compare across other geographic areas to rule out other factors. Another option I know if is doing periodic surveys with your target demographic before and after campaigns, to see how it changes over time.
posted by !Jim at 8:29 PM on March 25 [1 favorite]


« Older I was traveling today when I p...   |  My dad (who was Dutch) used to... Newer »

You are not logged in, either login or create an account to post comments