SEO

Large Site SEO: How to Waste Your #1 Ranking with Bad Formulas

If you’re a serious SEO, this is the post to show it in the ‘Comments’ below. The following example is a real world example, from a large site, and not for the faint of heart. Keep in mind that we’re looking for tips that work in the real world, not Utopia.

If you work on a large site, or plan to, you had better be pretty good at creating formulas that work across millions of pages. There will not be any one person writing millions of meta descriptions by hand for you. Admittedly, this can be a huge challenge and potentially leave some serious egg on your face when you miss the mark. You may end up ranking #1 for a key term and find that traffic just doesn’t come the way it should.

Here are a few examples of a site that ranks very well for <Geo> <Category> searches. Please note the meta descriptions, as we’re going to be tearing this formulaic approach apart in the ‘Comments’:

clip image002 0018 Large Site SEO: How to Waste Your #1 Ranking with Bad Formulas

Does this make sense? I’ll accept that ‘Plumbers in Pittsburgh maps’ is proper English (not really), but do people searching for plumbers need a map?

“Hello, Mr. Plumber, my toilet is clogged. I’ll be dropping it off around noon. Good thing that map will get me there.”

Now check these out:

clip image004 0006 Large Site SEO: How to Waste Your #1 Ranking with Bad Formulas

clip image006 0005 Large Site SEO: How to Waste Your #1 Ranking with Bad Formulas

At this point, from the three descriptions shown above, we can clearly see the formula applied is:

Directory of <City><Category> in <state abbv> yellow pages. Find <Category> in <City> maps with reviews, websites, phone numbers, addresses, and business profiles.

Now that we have torn apart their formula, I’m left with all kinds of questions:

Does this formula make logical sense?

Is it worth taking the time to address the fact that one formula for ‘Restaurants’, ‘Plumbers’ and ‘Lawyers’ may not be the right way to optimize click through individually?

What internal challenges might we be missing from the outside that explain this odd sounding formula?

Does the meta description make you feel that this site will have what your <Geo><Category> search was looking for?

Do you trust this site from what they have shown?

Along with your answers, my tips will appear in the ‘Comments’ below progressively. Let’s talk about why this approach works…or doesn’t. Let’s take an actionable tactic to apply to our own large site from this example.

Note: The site I’ve picked on here ranks damn well for almost every search I try. I’m using their formulas strictly as an example and in no way diminishing their SEO accomplishments.

 Large Site SEO: How to Waste Your #1 Ranking with Bad Formulas
Matt Leonard currently directs SEO, SEM and Revenue Management for Cruise Critic, the world’s largest cruise site and part of the Trip Advisor Media Group. You can follow Matt Leonard on Twitter to keep up with his updates. Feel free to ask about his latest charity project, ‘Tweet for the Cure’, to benefit Susan G. Komen for the Cure. The opinions expressed are that of Matt Leonard and not necessarily those of Expedia, Trip Advisor or Cruise Critic.

Comments are closed.

36 thoughts on “Large Site SEO: How to Waste Your #1 Ranking with Bad Formulas

  1. When I do find sites like this, though, I see it as a sign that there is a lot of opportunity to beat them at their game, if that was the niche I was going after anyway. Another good sign is when price comparison sites are getting top rankings for specific products. Just means no one has attacked that micro-niche yet.

  2. You’re absolutely right, Stephan. Beyond beating them, with many of these larger sites you can actually get more traffic even ranking below them just by having a more compelling search result.

  3. Hmmm… Well the formula in the meta description shouldn’t have any effect on rankings.

    As per Trust, I guess if I was looking for a directory of ideas of local businesses, this would be ok. But I agree that there should be a specific set of metas per industry.

  4. I had this problem on a site i was doing that needed formulas. It’s a job search site and each search (by job title, region, salary etc) created a different static search result page that appeared in the SERPs. I attacked this in two ways. First off i created a formula that worked as well as possible on as many pages as possible as the site in your post has tried to do. I then identified the most important 50 or so pages and did them manually. There’s no substitute to manually optimising tags, you just have to weigh up the time spent doing it versus the traffic you will miss by not doing it.

  5. You’re correct that the meta description alone shouldn’t have any effect on rankings. I’ve always operated on the premise that it does have an effect on click through, however, which in turn has an effect on the ranking.

  6. I believe that it would have an indirect effect insofar that click through rate probably *does* have a direct effect, and if your snippet is not as compelling as the ones beneath it, your click through rate will likely suffer, and you may find yourself losing ground.

  7. Ed Sexton says above:

    “I then identified the most important 50 or so pages and did them manually. There’s no substitute to manually optimising tags, you just have to weigh up the time spent doing it versus the traffic you will miss by not doing it.”

    This is totally correct in my opinion. It addresses two key points:

    1. All pages are not created equal. High value pages need a little extra TLC.

    2. Time spent vs lost traffic. We do not work for the company referenced in this example and have no idea what their resources, or priorities, are.

    With that being said, does anyone have a way to address both geos and categories across the board in a scalable, relevant way?

  8. Many Web sites employ these kinds of tactics. Working with large Web sites, there are sometimes many restrictions in their CMS.

    While the english and grammar isn’t perfect, as Jeff said this has no baring on actual ranking.

    I agree, it could help improve click-through rate. It is just hard to measure that organically.

  9. Using a compelling title and meta description per page is something most sites don’t seem to get.
    This creates opportunities to either grab those searchers or convince them to do hire a SEO company.

    In the above scenario I would advise something like:
    Title: Searching for in ?
    Meta description: All in the neighbourhood of are listed right here with all relevant information.

    The title is most important as lots of people don’t even read the meta description. The meta description shouldn’t be too long because Google can cut it short (which is not what we want) and when it’s shorter more people will read it. Thus creating more clickthroughs.

    I would furthermore suggest a test while leaving “with all relevant information.” away, to see what works better.

  10. Seems I shouldn’t use those halve aquares ;)

    I meant:
    In the above scenario I would advise something like:
    Title: Searching “category” for in “geo”?
    Meta description: All “category” in the neighbourhood of “geo” are listed right here with all relevant information.

  11. Thanks for the great comments.

    One thing I may consider doing with a directory site, that will work in scale, is sectioning services that go to your location (contractor) vs businesses that you go to (restaurant). I’d then give each section a unique formula that may be more relevant. In doing so, I may then consider removing words like ‘maps’ and ‘addresses’ for service oriented businesses where they may not be a compelling use of characters.

    So, I’d say at least one possibility for improvement lies in identifying the characteristics and applying different formulas for ‘things you go to’ and ‘things that come to you’.

  12. The one reason why this WORKS is that I think subsconsiously people are looking at the site snippet for the maximum number of words that map to their search terms (which undoubtedly includes “Pittsburgh” and “Plumbers”). A Meta Description that reads more like a real sentence might not map as well to the search query and actually generate less CTR.

  13. Good point, Todd. They do an excellent job here of getting their optimal keywords bolded twice.

    Perhaps, with the Plumber example, this works and accomplishes what you speak of:

    Directory of Pittsburgh Plumbers in PA yellow pages. Find licensed (master, 24 hour…) plumbers in Pittsburgh with reviews, websites, phone numbers and business profiles.

    What I did was replace the unnecessary ‘maps’ and ‘addresses’ with a more relevant term like licensed (which of course depends if the company in question verifies licenses for advertisers where required). It also leaves the option to insert ‘master plumbers, 24 hr plumbers, or another descriptive word of choice.

    The key here would be committing to all ‘Geo Plumbers’ falling under one formula (or possibly all licensed contractors). The Geo is the easily scaled part, in my opinion.

  14. Great article. SEJ quality post. This place rox.

    Do you need to take precautions to keep your small sites separate from the behemoth in the eyes of the SE’s.

    For example, would I not want to use the same adsense account on the small site as the big site? Would I want the registrar info to be different? etc…

  15. Matt this brings up a really important point – if we’re going to be successful on mass volume sites, we need to think through every aspect of on-site optimization and how we can automate every one of them but only to the point where we don’t then pollute the content, let alone the text that appears in the SERPs.

    When I work on this type of site, one of the things I do is check the actual phrase usage from city to city – quite often there is some variation so for example, someone looking for plumbers in Pittsburgh might use the phrase “Pittsburgh Plumbers” while someone in Las Vegas might use “Pittsburgh Plumber” or “Plumber in Pittsburgh”.

    While the automation of that much complexity is usually not worth it for every geo location a mega site covers, like Ed described for his “top 50″, I at least want to hyper-customize the top landing pages.

    One site we’re doing this on now has the top 100 cities targeted for customization. We’re taking it to the extreme and will have a completely separate section devoted to “Featured Cities”, and within that full on listing pages that also have supporting hand optimized content (text, photos, in-content links, etc. related to that city.

    We then also target those specific pages with quality inbound links.

    If our keyword research shows that there are two, or three or even four variations that get the most volume across the range of categories regardless of the specific geo target, that means we also have the opportunity to seed each of those at least in some aspect of the overall template structure of the site.

    That way we get at least some top ranking, but also come up in the 1st page SERP mix for phrase variations, and that leads to many more long tail visits as well.

    Oh – and let’s not forget about tailoring a custom PPC landing page for each of those!

  16. You are all throwing out great tips. Thank you so much for doing so.

    Personally, when dealing with a giant directory site like an IYP, I would start by identifying key categories. My belief is it’s much easier to scale geos than categories in an intelligent, formulaic way.

    For example, a formula makes more sense for use with Philadelphia Carpenters and Los Angeles Carpenters than Philadelphia Restaurants and Philadelphia Carpenters.

    Simply put, the easy variable is the geo and the difficult variable is the category. By grouping the categories as much as possible, with the criteria of your choice, you will have an opportunity to get your formula closer to perfect. The more granular you’re willing to go by category, the more your formula will convert.

    So, even if you have millions of geos, you’ll still only have hundreds or thousands of categories. If your formulas are isolated by category, manual formulas can be scalable.

    It would also be a priority of mine to make sure each category has a search result that is as compelling as any niche directory site competing in that particular category/vertical.

  17. Hi,

    The Title formula is ok and very recommended, You should make different formula for the description, make different formula for each and every category (locksmith, restaurant….) so the visitors will get text the define more about what they looking for…

  18. These formulas/logic (how they have their content and meta data set-up) usually has much less to do with how they rank comparable to how their internal linking structure is set-up. I would look at their internal linking for reasons for rankings.

  19. “The site I’ve picked on here ranks damn well for almost every search I try.”

    If it ranks well, and it’s so large, then why would you spend money and time manually writing thousands of meta tags?

    It’s still a business….

  20. Questions for the group….

    On a site as large as an IYP (10-20 million pages?), is it worthwhile to spend tons of time to micromanage the meta descriptions?

    How many full time people would it take to micromanage user experience every city and category available?

    How much will that effort improve the click-through ratios?

    How would you A/B test on Google without deteriorating the current rankings?

  21. Matt,

    @Jeff
    “Hmmm… Well the formula in the meta description shouldn’t have any effect on rankings.”

    completely wrong. (why people should do their own testing instead of reading and repeating seo blogs unless you know the person does their homework… i am reading matt’s article.. so that should say something about matt.)

    anyways… your premise is correct.. i have 14 sites o’m currently working on that this applies to.. and 30+ that it did apply to, smallest one has just over 8000 entries each with unique properties.. or DB criteria.. if u remember from #imsb database theory.. try and use absolute urls then your formula is attributed on a unique url basis..

    With larger sites.. say 50k – 500,000 entries in database.. the formula you speak of works…

    usually [subject] [phys address] [city] [county] [state] [zipcode] is a given… this allows for unique meta descriptions/titles/alts which fixes errors in google webmaster tools by providing unique descriptions for each page. IIS or apache can use rewrite mods to include it in the url too..

    yes i am talking about the highly competitive real estate market.. but a national plumbers directory is kinda the same thing.

    better semantics can provide better rankings, staying away from more than 3 KW items in a series helps as well..

    anyways.. your premise is correct (based on my extensive testing.. )

    good article… =)

  22. leroy j.. yes.. AND it’s billable.. =) BUT… with what matt’s talking about above.. you code it all and parse DB or SQL varibles so u don’t have to code each single page.. usually because anything over 1000 pages usually is in a CMS..

  23. Conrad: Because a site ranks well doesn’t mean there’s not room to improve. A proper meta desc can convert better, lower bounce rate by more accurately describing the experience, and improve CTR. Using your logic of not improving the search result, I’d compare that to writing a PPC ad and just being glad it appears. Why worry about whether it actually works or not? It’s still a business, which is exactly why you optimize everything. Contrary to what you seem to think about the resources required to do this, it’s actually pretty easy and not on-going.

    Steve: Thank you very much. I do know how much you test, and go against convention successfully, so I really appreciate your words.

    Leroy:
    On a site as large as an IYP (10-20 million pages?), is it worthwhile to spend tons of time to micromanage the meta descriptions?

    It’s actually not tons of time. From what I know of the IYP world, there are probably 200 categories at most that account for the vast majority of importance and dollars. Writing a formula for each category (top 200) by hand is tedious, but I think it’s one week worth of work, maximum. Remember, the millions of pages come from the geos, not the categories. And, the geos are the easy part to plug into any IYP formula for title or meta desc.

    How many full time people would it take to micromanage user experience every city and category available?

    See previous answer. 1 person, 1 week. Optimize formulas for 200 categories (or # of your choosing) and the geos fit easily.

    How much will that effort improve the click-through ratios?

    Depending on how bad the current formula used is, upwards of 20%.

    How would you A/B test on Google without deteriorating the current rankings?

    Really a bunch of ways here. PPC is the most measurable and fast. The key is to find the right buzz words (licensed, affordable…whatever fits) and test them in different ads. Your PPC ad effectiveness can very much mirror a portion of your meta description. Keep in mind, that probably half of the meta formula will be the same across every category for an IYP regardless. It’s the other half that will help. You can also split test with generic formulas vs optimized formulas and measure growth, but thats much more of a moving target.

  24. I once had a website back in 2005 where I had 64,000 pages, each representing a particular city in the United States. I employed similar tactics of dynamically changing the meta tags and content for each page and that worked great for a while especially with Yahoo sending tons of traffic my way. But it barely lasted 6 months before the traffic stopped, guess the search engines recognised there wasn’t too much unique content on offer (used customised rss feeds and basic database info on each city). If the content offered on each page is substantially unique and informative, I think it should work well – personally, don’t think meta tags will hold that much weight.

  25. Well in addition to conversion, I feel Title is the most important & first aspect which grabs user attention and lead to more click through organically. Search Engine’s too pick the to display in SERP’s as it is as given on site.

    While with many time’s I experienced Search Engine’s do not display the exact description what we provide over site at SERP’s. They sometimes pick snippet text from Dmoz description or somewhere within from the page content. More weightage has to be given on formulating title.

    I feel the description Formula used here is perfect and having City & Category bold in description will help both way Organic Rankings & Clicks.

    I am doing SEO for a very huge site where we are targeting almost every city of US for our products and services. It was tough for us to write meta’s for each and every webpage unique, so we also created the Formula to be used with CMS as below:-

    Title:
    City Category | Category City, – Domain Name

    Description:
    City Category – XYZ provides Category (Product & Services) in…text to complete meaningful sentence…. City (Abbv).

    It has worked great for us, we are almost on Google 1st Page for most of our targeted terms (many other Onpage & Offpage SEO factors do count), even Clicks are aussum for some of our listings at 7th & 8th position.

    However I am looking forward to learn a lot from this discussion, waiting to see what all are will the outcome’s of this discussion.

    Cheer’s

  26. SEOMOZ.org discuss GEO search with regards to consistant address details across the web with map details and placing that information on informative sites relevant to that location. i.e. listing your site with the local council website or chamber of commerce site. Search Engines will dictate relevance based on how consistant your addressing is across the web. These formulas could become handy in other forms not just for meta tag descriptions etc!

  27. Dear i see your sites and i like your sites very much but i request that share more knowledge about SEO, SMO, and on page and off page also. Hey guys my company websites information is here which i want to share with you and also need response from you that what about my sites. My sites information is here:-SEO Bedrijf :For SEO Services India visit Profit By Search A SEO Company in India offering SEO Services, SEO Consultancy Services to Companies in North America & Europe.

    Thanks and Best Regards
    Tauqeer Aziz

  28. Dear i see your sites and i like your sites very much but i request that share more knowledge about SEO, SMO, and on page and off page also. Hey guys my company websites information is here which i want to share with you and also need response from you that what about my sites. My sites information is here:-SEO :Learn how search engine marketing, PPC management, conversion enhancement and web analytics tracking can boost your online profitability. Visit Alkemi’s website for more details about Internet Marketing Services.

    Thanks and Best Regards
    Sam

  29. George Jessup was on duty, driving his plumbing company’s truck. But when he saw a duck sitting on a storm drain on a busy road in Shaler, he didn’t think twice about stopping.

    “The duck wouldn’t move, and it was in a dangerous position on Anderson Road. It piqued my curiosity,” Mr. Jessup said.

    When he approached the duck he heard “peep peep peeping” and realized there were ducklings in the storm drain. He just had to help.

    Mr. Jessup, 49, of Millvale, got busy on his cell phone. His first call was to his boss, Bob Beall, franchise owner of Mr. Rooter Plumbing of Pittsburgh. He told him to go ahead and help the ducklings, “and he said he’d cover me” if customers called with plumbing emergencies, Mr. Jessup said.

    When he couldn’t budge the heavy grate, he called Shaler police and volunteer firefighters. The ducklings “were flipping out,” he said, and the mother duck, a mallard, was agitated.

    “The firemen took the grate off. It must have weighed 300 pounds,” Mr. Jessup said. “Then a police officer reached in and picked up the ducklings. There were six or seven. They were the cutest things you ever saw — brownish with some white and yellow.

    To make sure they got every duckling, Mr. Jessup sent Mr. Rooter’s video inspection camera down the pipe. There were no more.

    “It was just inspiring how everyone pitched in,” he said.

    After their three-hour ordeal, mamma and her ducklings “just waddled off,” he said.

    Mr. Beall got one emergency call while Mr. Jessup was rescuing ducklings; he called another plumber in on his day off.

    “Mr. Rooter practices courtesy as a normal part of doing business,” he deadpanned.

    Call (724) 716-4099 for the best plumber in Pittsburgh!

  30. This is a fantastic post! An absolute eye-opener. I never noticed that such use of meta description is also done. Everybody’s feedback is also worth reading. This is something one should divert their attention too and take into notice. And I too think that using such meta description will not help such sites to grow their business opportunities. A must read post by Matt Leonard.