23 replies to this topic

[qwerty]

I guess I'm kind of in the middle on this. I agree with Chris that rankings aren't really meaningful simply because you can't know what other people see when they run a search. That's why Webmaster Tools shows you "average" rankings instead of actual rankings, even when you're looking at data gathered over a short period of time.

But I also happen to be employed by a global network of very large sites (very) that targets a lot of different keyword phrases, and we have seen that higher rankings (even average rankings) = higher click-through rates. Moving from page 2 to page 1 makes a difference, and moving from the bottom of page 1 to the top makes a bigger difference. So yes, we track rankings, we generate reports, and we make decisions based on those reports, then track the rankings and generate the reports again to see if those decisions made a difference in rankings, traffic, and a number of different forms of conversions.

We used Searchlight in the past, and we currently have Authority Labs. Both are pretty expensive, but that's at least in part a function of how many keywords we track (currently over 60,000 across numerous domains, languages, etc.).

[tcolling]

Thanks to all for taking time to respond to my question. I appreciate your help!

- Tim

[Tiggerito]

How is checking rankings going to tell you anything about that?
You see what you see which may or may not be what others see.

Nope Alexa figures are based on the less than 4% of Internet users with their toolbar installed then extrapolated, that's a 96% margin of error.

So you should completely ignore both?

Yesterday I took the top 100 SERPs for around 1,500 results in a clients business. I calculated estimated click rates for each entry and added them up to score each domain. From that I picked out the best scoring domains that are worth checking out for link opportunities. Is this a complete waste of time because SERPs vary and there is a margin of error?

What I'm saying is the data is still valuable but you have to understand its limitations. I know that my results are skewed because of location, time, being anonymous, the size of results requested, traffic data errors, formula errors and that I can't do massive long-tail analysis. But it still has value. My client now knows which domains have value in their market.

Because of the reasons you stated I don't report on individual keyword progress unless specifically asked to by the client. I often have to re-explain to them that they are not #1 outside of their own browser!

p.s. that's not how you calculate margin of error

[qwerty]

In the case of Alexa data, yes. I think you should ignore it completely. The fact that the data only comes from a small group of people is only part of the issue. If that small group represented a cross-section of everyone on the internet, then it might mean something, but that's not the case. People with the Alexa toolbar installed are almost invariably going to be people who care about the kind of data Alexa provides. It's a pretty specific group.

If your market research is based exclusively on surveys of marketers, you're not going to learn anything from it unless you intend to market whatever it is you're marketing exclusively to marketers.

Edited by qwerty, 30 June 2012 - 11:38 AM.

[Jill]

Tiggerito, what you're talking about is different. Doing competitive analysis that way is very valuable. I'll often look at the top 10 or so pages that show up for various phrases that my client may think they should be showing up. Mostly it's to explain to them why they don't (because those other sites are spending tons of money and are way better than them!).

But none of that is the same as running ranking reports and clients thinking that things are good or bad because of them.

That said, there is some value in spot checking some rankings at some times. Like when you see in your analytics that certain phrases have lost (or gained) traffic, you can usually confirm why when you do a quick search to see approximately where they're showing up. If they're not getting traffic for the keyword anymore and used to be in the top 10 but now aren't you've got to figure out why.

[chrishirst]

p.s. that's not how you calculate margin of error

Yes I know how to calculate a statistical margin or error, I was not trying to be literal, Mainly because a TRUE margin of error requires the sample numbers to be known.

Let's say then that In the case of Alexa, by extrapolating data garnered from ~4% to all Internet users leaves a worst case scenario of getting it wrong ninety six times out of one hundred.
Even more so when you know that ~75% of Alexa Toolbar users are located in South East Asia.

My client now knows which domains have value in their market.

Ermmmm? Domains?? No. Search Engines rank URLS not "domains"

I calculated estimated click rates for each entry and added them up to score each domain.

How???

[Tiggerito]

@chrishirst & @querty I totally agree with all of you on Alexa with their South East Asia marketeer based data.

@Jill, I spot check as well. Recently I've done some indetail analysis of specific search results to find actionable items. I think people forget the results are for more than the ten blue links.

Ermmmm? Domains?? No. Search Engines rank URLS not "domains"

I group my results into domains to make the data more usable. Quite often a single domain does well for a series of URLs, each targeting different keywords in the market. Finding those top domains then digging deeper is the general process. One trick I found was that I can discover the best URLs for the top performing directories and then place my client on those pages.

How???

If I tell you I will have to kill you.

I've actually created an online tool that shows how I parse SERPs. Just google "Google Search Result Parser" to find it (hopefully).

For each parsed component I estimate clicks per search based on the Chitika data combined with my own alterations related to the different types of components in search. e.g. an organic result with author and review stars will be considered more likely to be clicked on than a mini-sitelink.

This is then combined with traffic estimates (AdWords) to come up with an estimated CTR. I add up the CTRs for each URL and domain to get overall scores.

There's more to it, but that's the basic idea.

Lot's of room for error and don't take the CTRs as being even close to the truth. It's just a way to help me sort big data so I can spot and focus on what's probably most valuable.

Looks like they've changed the star rating code on me again :-(

[Hex]

I was pointed to this thread from another where the same question was asked.

I was a little worried to ask this question as I knew that this would be the answer but for those of us who do need to track this would anyone be able to recommend a good looking tool that can produce nice reports?

I totally understand personalised search and rankings, ROI and all the rest of it but I do need to track the rankings, its part of my job and takes up a lot of time. If I could track rankings then it would save me ages and I would be able to focus on other areas.

[Jill]

Hex, please read through the thread as there are recommended tools that have already been posted throughout it.