• Richard Thomas

    Hi Sam – great post – can you elaborate a little more on why Low Search Volume can be damaging?

    • Sam Owen

      Hi Richard – good question! I think I might have phrased that a little awkwardly. While there is some debate as to whether Low Search Volume keywords are actually bad for things like account level QS (I don’t think they are) – they do cause clutter and confusion, which are generally bad for an account. No need to panic and go and remove all your LSV keywords though.

  • Chris_LaRoche

    Great post, Sam. This is something that I’ve thought about a lot. Usually I find myself starting out with a monster excel build and end up having to trim back later. I think for my next new account build, i’ll try a more streamlined approach.

    • Sam Owen

      Hi Chris,

      That’s exactly how I’ve always done it – monster Excel build and trim down. The meat of my post is that I think we should start doing it the other way round. Still a firm believer in keeping in LT keywords that are getting clicks/conversions.

      • Chris_LaRoche

        Definitely. I am on board. I’ve got a new account launching soon, and I think i’m going to take this more streamlined initial approach and add in any longer-tail after I see them converting for us.

  • Martyn Wright

    While I agree with the logic of having long tail keywords in the account for relevancy, this is always the issue that crops up. The (modified) broad match picks up thousands of variants, many of which convert, but individual conversions are typically just that, with that query never converting again. By adding them to the account you soon end up with thousands of long tail keywords with no traffic volume that Google refuses to show due to “Low Search Volume”. Go too broad and you hit the low quality score due to not being relevant enough, so finding the right balance of a keyword with enough “long-tail-ness” to be highly relevant, while also being broad enough to have volume can often be quite tricky.

    • Sam Owen

      Good points Martyn. That’s why I tend not to add a LT keyword to my Top Performer campaigns until it has at least 2 separate conversions.

  • http://fat-karate-ka.com Stacy

    Long keyphrases are no longer working for us, at all. All of our long tails, and I’m considering long tails as 4+ words, were doing a great job for us until about a year to year-and-a-half ago. They were inexpenive, we were getting people into very specific product pages and we were doing great conversions. Then Google marked them as “rarely searched” and stopped showing them. I kept hold of them for a while but, after months of them not showing, I started to pause and delete those so I could more easily visualized what was showing and working. I understand that Google wasn’t making a lot of money on the long tail keyphrases, but I feel like my companies were penalized because the others in my industry were too lazy to set up an account structure that benefitted searchers.

    • Sam Owen

      The theory is that Low Search Volume keywords should still work if people actually search for those terms. However I’d love for you to pull some numbers for me – download a report of all your LSV keywords for say the past 30 days and tell me clicks, impressions, conversions. Then take those same keywords and take a random 30 day sample from say 1-2 years ago. Were those keywords generating many more clicks/impressions back then? If so you have a case to say that LSV keywords are stopping ads from showing. Otherwise it might just be that account changes / high bids on broad terms etc are taking over. Thanks for a great point!

      • http://www.webranking.com/ James Svoboda

        I’ve seen the same pattern with Low Search Volume (LSV) keywords. Most were active a year or two ago, had decent impression volume, but then just dropped off the face of search. IMHO AdWords has the LSV dial tuned way too far up.

  • http://www.webranking.com/ James Svoboda

    Hi Sam,

    Good post. I’m a firm believer that long-tail keywords, keyword with more words in any given keyword, can greatly define intent and increase conversions. I gave a presentation at SearchFest a couple of weeks ago where I sent more than a few minuted going over this.

    http://www.webranking.com/blog/creating-linear-models-that-lead-to-successful-ppc-campaigns

    I could tell that some didn’t quite fully comprehend how this relates to various aspects of a PPC campaign, so it is nice to see that another understands the implications.

    • Sam Owen

      Thanks James!

      Just took a look through your deck and agree esp. with tiered bidding. Getting intent right definitely can be tricky with anything other than exact matches, but a smartly laid out Mod Broad set up (like in your deck) can be 99% effective.

  • Andy Brice

    Sam,

    A very interesting analysis and some creative use of pivot tables. Great to see lots of real numbers. The exact definition of long tail is, of course, open to interpretation (every bit of a power law curve looks the same as the whole). I have seen plenty of Adwords accounts (set up by non-professionals) with 1000 keyword phrases as long tail!

    I guess it all comes down to cost (time to add keywords) vs benefit (increased profit). I have just released a new product aimed specifically at helping Adwords users add large numbers of long tail keywords to their campaigns. This should significanly decrease the cost side of the equation compared to using, say, Excel. Excel is a marvellous piece of software, but it wasn’t designed with Adwords in mind! If you are interested, please check out http://www.keywordfunnel.com .

    • Sam Owen

      Thanks Andy!

      I agree with you about the definition of the long tail being open to interpretation. There’s also a lot of confusion about long tail search queries vs. Keywords. I hope no one reading thinks I’m throwing out long tail search queries as being important (pull a pivot table and see for your account) – but rather the inefficiencies of current account structures to accurately capture them & the difficulty in optimizing such huge behemoth accounts (Can you really test ads across a million ad groups?). I’ll check out the link!

      • Andy Brice

        My own approach has been to aim for 10-20 keyword phrases per adgroup. Then set the the bid by adgroup using CPA. This makes campaigns with thousands of keyword phrases much more manageable. But you need to keep an eye on ‘dimensions>search terms’ for negative keywords. Also CPA bidding isn’t appropriate in all cases.

  • RIna Ashton

    I’ve
    spoken with a lot of PPC companies, but the most important thing that

    you can actually do to your marketing campaign is landing pages. Try out

    this company results driven, call them at 734-344-6781 . They are
    by far the

    best marketing company to go with since they not only create solid ppc

    campaigns for you, but they also do landing pages, CRO and remarketing

    which are all 100% necessary elements of a successful online marketing

    campaign. Most companies sell you 1 single thing such as CRO, or just PPC,

    or just popups, etc. This is completely ludicrous because while 1 thing may

    make a difference in a stable marketing campaign, there is no single

    element that is a make or break element in online marketing, you need the

    entire package, and then hone/optimize from there.

  • Steve Cameron

    I’ve wrestled with the time vs. results issue with respect of LTK and have generally come to the conculsion that the work is simply not worth the return in most cases. Of course, it does depend on the industry and LT can be very effective, for example, in a competitive real estate market where the return from the sale is substantial. For sectors with smaller unit costs it’s hard to make it work.

    • Sam Owen

      Thanks for the comment Steve and you are 100% right about it being industry specific – we have a client in the real estate business and it is mostly about long tail for them with specific MLS numbers doing really well in aggregate.

  • Amish Zinzuwadia

    Brilliant! I have been thinking like this for a while and it’s good to see a case study to highlight how we’ve been blindly led down the path to expand infinitely

    • Sam Owen

      Thanks Amish!

  • Heath Showman

    Great article Sam. Can you clarify whether in the new Top Performers campaign you create an exact match only keyword or a broad match too?

    • Sam Owen

      Hi Heath,

      Top Performers is for exact match single keywords ad groups only. The broad matches just remain in the general “mining” campaigns.

  • Ben

    Hi Sam,

    When building a new account structure, you wouldn’t know which keywords are the top performers. So how could you you model this approach when building out an account?

    • Sam Owen

      By making liberal use of modified broad keywords in your “general” campaign. 4-5 word long modified broad keywords will pick up all of your long tail search queries. Once those MBM keywords have picked up the SQ data, you can spin those long search queries into top performer exact matches.

  • Steve Squier

    Can’t play too conservatively when you’re doing PPC. You have to bid to win that traffic. You should be receiving >90% impression share of the keywords you are bidding on and that means bidding a little higher sometimes. What happens is that you have to have an optimized sales process to justify the ability to bid higher. If you make more per sale, you can bid more per click, and squash your competition. My buddy Simon would be willing to help anybody that wants help with their PPC campaigns, just give him a call. His number is 325-446-1507 .

  • Akash Agarwal

    Nice article. I never use long tail keyword in ppc.Thanks a lot for sharing such kind of information.

    • Sam Owen

      Thanks Akash!

  • Brodie Tyler

    I feel like 2-4 (maybe “mid-tail”) keywords are the way to go. This might just be projection and/or my own arbitrary preference, but I have an aversion to single keywords because they seem insufficient ly descriptive, while anything longer than 4 at the most, seem unnecessary, distracting and somehow pretentious. Again, could just be my own strange neurosis.
    Brodie

  • Ken Fobert

    Great article Sam. Couple questions for you in terms of how you structure your campaigns:

    1. In your general campaigns, I’m assuming you have multiple keywords per adgroup with all 3 match types? Do you structure these adgroups based more so on a theme (i.e. keywords in 1 adgroup are similar but not exactly the same) or do you separate them by keyword (i.e. each keyword gets it’s own adgroup with modified broad, phrase and exact match types)?

    2. In your top performers campaigns, you say each keyword gets its own adgroup. Do you move over the exact search term that triggered the ad (and 2+ conversions) whether or not its a broad, phrase or exact match keyword? For example, using your hypothetical company if you saw that a phrase match keyword for ‘ppc management agency’ triggered at least 2 conversions in the last week do you just move that phrase match term to the top performers campaign, or do you add the modified broad, phrase ad exact matches of that phrase as well?

    Once again great article and you’ve given me some new pivot table ideas that will come in very handy.

    • Sam Owen

      Hi Ken,

      1. In the general campaigns we have a mix of match types in ad groups organized around tight themes (normally 5-10 keywords per ad group) – sometimes we will break out by match type but this often depends on the account.

      2. We move the search query over as an exact match only. If a broad match has 2 conversions, but for two totally different queries, it won’t get moved over into the top performers campaign. We want our top performers campaign to reflect the terms people are actually searching for.

      Thanks for reading!