• http://www.rimmkaufman.com/ George Michie

    Fred, I love you man, but I’d take exception on a couple of points:

    1) with 2 clicks on exact match “Belgian chocolates” you don’t have enough data to conclude that that version doesn’t work well. Indeed, many many years in the business suggests that the exact match version will perform better than the broad match, it simply isn’t getting any traffic right now because you’re bidding the same for all match types;

    2) bid to first page min is almost always an ROI negative proposition. We also built the ability to reach for it if we’re close, but find that our bidding system is generally right: if the traffic is worth at max $0.50, bidding $0.55 to get traffic just gets you traffic that isn’t worth the cost. As part of a portfolio that may be better than other bad traffic to buy, but…

  • http://www.toptiertools.com/ Frederick Vallaeys

    Hi George, love right back at you! I always like hearing from smart PPC people!

    1) Totally agreed that you always need enough data before making a decision and 2 impressions is never enough data. I should have selected a better screenshot to illustrate that… The overall point is that Impression Share at the keyword match type level gives us stats about a more specific targeting criteria than what we had previously so now we can make decisions about smaller segments of our account and that allows us to extract more value from AdWords than before.

    2) If you have sophisticated bid management in place, then you’d certainly leave the bidding to that system. But when you’re bidding manually like many advertisers, you may find hidden gems that never had a bid that was high enough to get any meaningful data (which as you pointed out you need to make smart decisions). If you are bidding manually and you relegated a keyword to page 2 because it wasn’t profitable on page 1, you’d also not want to raise the bid again. That said, I believe most advertisers who are stuck on page 2 because page 1 is ROI-negative should not think of the cost of AdWords as the problem. Instead, they should look for ways to improve their landing pages for better conversion or fix their business so that their margins are competitive with the advertisers who are somehow able to afford page 1.

  • http://www.rimmkaufman.com/ George Michie

    Good points, Fred. As with all things the advice we give is colored heavily by the types of clients we work with and what is great advice for a local B to B service company may be bad advice for a major national retailer. We should all have to fill out a grid when we write posts or comments that specify the types of advertisers (vertical, business model, geographic reach, size of spend, etc) for which the advice makes sense to avoid misunderstandings. Rock on, sir.

  • Pat Grady

    “you can make changes for that one keyword without worrying about impacting the other keywords in the ad group.”
    Articles do have space limits, I’m gonna let you get away with this (yes, I’m in charge).

  • ChadB

    Forgive me if I’m wrong but I believe you are misinterpreting what Impression Share can tell you.

    From AdWords:
    Search Impression share is the impressions you’ve received on the Search Network divided by the estimated number of impressions you were eligible to receive.

    The details: How often your ad is shown depends on your ads’ targeting settings, budget, approval statuses, bids, and Quality Scores.

    That last bit about Quality Score is the issue.

    Since we have no idea how AdWords distributes (by percentage) impressions based on QS it is impossible to know how much of the market we are actually missing. We only know that based on OUR quality score we could get more or less IS.

    Google of course does this so we can’t know exactly how many searches are being performed on their networks for specific terms. Unfortunately, when clients as how many impressions are available, as a whole, it’s impossible to know.

    You could say we should always work on improving QS, but that’s not always possible based on factors that are out of our control. For example, I had a client with a low quality score, I called AdWords, and was told the company was categorized in a product category that the company was not a part of and not selling into. As a result the cohort was incorrect, the company wasn’t getting as many clicks as “expected” and was getting a low quality score.