[...] whose properties are well-optimized for organic search (per studies by Andrew Shotland and Ash Nallawalla), the 10-pack has slowed referral traffic from Google considerably, and even Google Suggest thinks [...]

[...] Nallawalla, one of two Australian Local SEO’s that I know, authored an excellent followup shortly afterwards, focusing on a broader range of geographic areas, but a smaller number of [...]

[...] Ash Nallawalla/Net Magellan: SEO ranking of US IYPs across 274 cities [...]

[...] Ash Nallawalla/Net Magellan: SEO ranking of US IYPs across 274 cities [...]

[...] « SEO ranking of US IYPs across 274 cities [...]

Google Trends gives me a different result for the current date – it ranks Yelp, Yellowpages.com, Citysearch, Superpages and Insiderpages. But it’s just as unreliable as the others, as the trend line for visits to our site doesn’t resemble our own data (which includes direct and non-G traffic).

Fair point on the state abbreviation, but I was looking for just one consistent keyphrase across the country for my IBP evaluation. If this were a paid assignment, I’d partner with one of the US SEO companies that have some powerful black-ops ranking checkers distributed over dozens of IP ranges and a decent database. Then we could dig up some realistic keyphrases across the more lucrative headings.

It’s interesting when Yellowbook powers IAF but their rankings are not adjacent all the time.

I haven’t heard of IYPs selling the SEO sizzle as part of the offering – certainly would make me nervous if they did.

Both you and Andrew are demonstrating an approach that I also have used many times in assessing “how well we are doing” in terms of SEO for a particular site. While individual keyword rankings are not the only measure, they are one good way of seeing if one is obtaining the results one is seeking when doing SEO.

While I applaud your methodology, I differ with a few of your conclusions.

You’ve noted the limited scope of this ranking test, and, just as you point out, it’s not particularly practical to try to discover and assess rankings for every possible keyword phrase combination that users would type into search engines when seeking YP type information. However, just as you did, and just as Andrew did, I think it is feasible to assemble a sample set of queries that’s representative of what typical internet users will type into search engines, and project overall results based on the sample set.

To do that, you need to form your queries as the majority of internet users are forming theirs. Your sample format appears to be: “city + state-abbrev. + category”. However, most American format their queries without the state abbreviation, so Andrew’s sample queries would actually be a little bit more representative of the majority of consumer queries, in my opinion. (There are exceptions to this, of course, such as cases where Google does not easily identify which city for cases where there are many cities sharing the same name such as “Springfield”, or particular cases where regional users settle upon an alternate name such as how people in Manhattan use “NYC” at least as often as “New York” or “New York City”. But, in most cases, internet users are typing in standard city names sans State name qualifier — a practice growing even more common as Google is defaulting geotargeted Maps results more and more.)

There are very definitely instances where IYPs which are performing well in a broad sample set are not performing so hot for smaller markets or particular regions where their competitors enjoy “incumbent” status as the main/traditional print directory provider. This is an area that makes comparative analysis extremely hard to do, because you could have a site which is ranking very well for most of the major metros, while having zero presence in the thousands of small towns out there. Unfortunately, comparing rankings for a search term in a small town versus major metro is not an apples-to-apples comparison for a few reasons. First, small towns do not all have nearly as many types of businesses, so while a big city might have a “divorce lawyer”, many small towns will not. In those cases, an IYP site which has a page ranking for that small-town query may be wasting some huge amounts of PageRank in category+metro combinations where they have no real content. Secondly, ranking well in one major metro for a category could equate in terms of traffic volue with ranking for the same category in hundreds of small towns.

So, while we might like to get a more comprehensive sample set that encompasses thousands of locations combined with hundreds of YP categories, it’s fairly safe to project that those IYPs that are consistently ranking well across the top metro areas by top popular YP categories are likely performing the best overall.

Your article has another salient point when you mention that some categories are worth a lot more to IYP companies than others, and perhaps some companies ignore those verticals which are worth less. So, categories for Post Offices and other government offices along with Churches and Non-Profit Organizations will tend to be nearly worthless in terms of potential revenue from prospective advertisers.

This is a perspective I run into with some large-company clients I have which are not YPs as well — they tend to be focused more upon the traffic that will convert. But, this is not a perspective I share without significant qualifications. In Paid Search Advertising, this is a major issue. In SEO, hyperfocus on revoking high-traffic content which doesn’t convert is a habit which can sabatage the overall success of a program. If you have content which validly ranks for some popular term combinations, and your site achieves good click-through referral traffic for it, this success can bring the all-mighty inbound links for that content and Google is also observing whether users stick in your site or rapidly bounce out — the good CTR within Google SERPs can help your overall site perform well on other search terms as well.

Finally, I’d exercise care about really trusting Compete.com, since their sampling can be highly skewed. They’re probably purchasing anonymous log traffic from a number of small sources such as ISPs and minor search engines, and assuming that their traffic will represent overall internet usage may be very unrealistic (rather like assuming that Alexa rankings are all that dependable or representative).

I think it’s better to check out Google Trends, since this shows relative rankings according to what Google themselves “see”. Example:

http://trends.google.com/websites?q=superpages.com%2Ccitysearch.com%2Cyelp.com%2CYellowPages.com%2Cinsiderpages.com&geo=US&date=ytd&sort=0

If we compare Andrew’s top ranking sites with Google Trends’ ranking of the same sites, we find that they coincide remarkably closely, if we aggregate/average the results for the past one-year period. (We have to remove Yahoo’s yellow page site from the list for this purpose, because 3rd-level domain results don’t appear to be segmented and reported separately for Google Trends.)

In any case, I can quibble with some of your specific methods, but the overall approach is good, and as you noted, there’s no simple way to rank IYPs. Using methods such as what you and Andrew demonstrated, combined with more robust 3rd-party analytics data such as comScore, Hitwise and Google Trends, all these various combined methods are necessary to try to assess how a site is doing in comparison to its competitors.

Thank you for going to the effort to provide such interesting audit results and meaty discussion content!

I trust the international version is coming soon?

Nico