Why Toolbar PageRank is Worthless

Soooo, a few readers asked me why I don't use PageRank as a statistic.

I thought about writing a long, angry, ranty post about Toolbar PageRank, bank bailouts and the moron on NPR today who said the USA has reduced its carbon emissions since 2000, etc. etc. but stopped myself.

Instead, I've got a nice clean list of well thought-out reasons:

Toolbar PageRank does not equal true PageRank. Toolbar PageRank tries to distill this humungous, universe-encompassing scoring system (true PageRank) into a 0-10 scale. Imagine buying a house based on a photo of a 1ft by 1ft square of green grass. Same thing.It updates infrequently. Actually, it updates slightly more often than the Department of Homeland Security's threat system ('High' since 2003, I think).Toolbar PageRank ignores onsite factors. OK, that may be an exaggeration. But that little green bar has little to do with relevance, and a lot to do with authority. It leaves out half of the SEO equation.It makes you focus on the stupid. Toolbar PageRank ignores your traffic, your sales, conversion rates, visit quality and the key phrases actually sending traffic your way. Aside from that and its total inaccuracy, it's perfecto.Back in 2008, Steve Reubel said that PageRank is the ultimate metric. He's also heralded the death of SEO at least three times. It's like a zombie movie, where the government tells everyone to go gather in a football stadium. Run. The other. Way.

So, if you want to slip on the proverbial banana peel. Otherwise, avoid PageRank and use some more reliable SEO metrics, k?

By the way, I wrote about PageRank and it's utter worthlessness in 2006, too. I also wrote a more succinct post on the subject in 2009.

comments (7) | trackbacks (0) | permalink

View the original article here

Market research is (often) worthless. Here's why.

Here's the short version: Knowing that someone has disposable income doesn't tell you what they want to spend it on.

I used to work for a (retail) company that had fairly deep pockets. Occasionally, we would reach into those pockets to buy market research.

Our database was scrupulously maintained, and could tell us basic stuff about our customers, such as name, address and complete purchase history. The market research companies we hired promised to expand on that data and provide rich demographic information about the folks in our database.

You can guess the next step. In strict adherence to the "birds of a feather flock together" model of marketing, the research company would then offer to sell us information (names and addresses, mostly) of people who were just like our best customers, with one distinguishing characteristic: they had not yet made a purchase from us.

Makes sense right? And especially tempting since, like nearly all retail businesses, we spent such a seemingly inordinate amount on trying to attract new customers.

Just one little problem: It didn't work.

For our tests, we came up with a very attractive offer and sent it out to three groups: Past customers; people who had a similar profile to our customers but had never purchased from us; and an equal number of randomly-selected households. All were within the same geographical area. Here's how it shook out:

The results were the same in three different trials, using three different profiling techniques.
I have mixed feelings about this. All thinking people should.

As a marketer, I was of course disappointed by the results. Everyone in business is always looking for the straightest and truest path to new customers. Intuitively, we are attracted to the concept that past behavior can reliably predict future behavior, and, to a degree, this is true. In our case, for example, we could reliably predict that people who were previous customers had a much higher likelihood of buying from us again, compared to folks who had never before made a purchase from us.

I'll say it again, with feeling: Knowing that someone has disposable income doesn't tell you what they want to spend it on.

What we couldn't seem to find, however, was an accurate profile of those individuals who, while not having yet shopped with us, still had a higher predisposition to do so.

The reality is that people are not that predictable regarding at least some activities in which they have never previously participated. In our case, just because someone had a profile that was in many ways very similar to that of our best customers, didn't mean he had any higher likelihood of buying high-end electronics.

Which, if you think about it, makes perfect sense in light of our everyday experience. Just because someone has disposable income doesn't tell you what they want to spend it on. Skiing is a relatively expensive activity. Yet we all know wealthy people who wouldn't dream of going to the slopes and poor students who somehow manipulate their budgets to indulge their hobby. It's the same with all sorts of non-commodity goods and services: wine, boats, concerts, travel...even relatively inexpensive activities such as movies or attending sporting events.

So, while as a marketer, I'm faced with a remaining challenge, as a member of the human race, I'm delighted that each of us is ultimately unpredictable - similar in many ways - but unique.

comments (6) | trackbacks (0) | permalink

View the original article here