This is a post borne mainly out of guilt.

Instinctively, without thinking, earlier on this evening I clicked a 'thumbs down' against Rand's video post about buckets of keywords. Of course, it means I have to write a post to explain my dastardly act of down thumbing, and having listened again to see if I was just being mean, my view is now that actually I probably don't know. Maybe.

But something did rankle; you decide whether this works or not, I'm not sure myself. 

Talking about buckets, Rand said the key seemed to be high traffic, good intent and high conversion rate. Well yeah. But of course, it never seems to work like that, and I've been thinking a lot about patterns and causes. Looking at your keyword research, I think the traffic levels can be where you can start to go wrong. Paid or natural. Simple stuff, but the broader the terms, the higher the traffic. They just look so darned attractive, don't they. Anything over a couple of thousand and the creative juices start to get going.

Conversely, the more specific and longer the phrase, the less traffic there tends to be. And all of your marketing instinct tells you to go for the high traffic, go for the mass market. Only, the more specific longer terms can be more valuable. By definition, these specific people are closer to "knowing what they want." Ergo, they are later in the buying cycle and worth getting to know, quickly. More specific searchers are also easier to pander to write for. You spend less time presenting a wide range of options, and honing the message is less of a balancing act.

So - a small point perhaps, but there are always less of these specific searchers, and you can never get enough of them to make or break a business, but they could be easy wins. Does this mean that instead of a couple of pages catering for very large numbers of broad searches and running tests ad infinitum, instead we need very large numbers of pages supporting fewer, but well targeted visitors?

Marketing used to be about reaching large numbers of people and persuading them to listen (2% a good result?). For me, it's tempting to start at the 2% and work up to the broader stuff. Feel free to disagree and go for everything :-)

Apologies if this has been picked over before. It's my first go.