Thoughts on How You Use Google Auction Insights

[Dane Saville]

It’s likely that most of you have seen the Google Auction Insights Report with new, crisp visuals. To our company, it’s most effectively used for requesting more budget and explaining KPI fluctuations, especially since you can segment by periods of time and see an account’s impression share trend over time.

So, if your impression share for specific keywords or queries decreases over time, it’s likely a sign of market volatility: (1) Your competitor’s could’ve started spending more aggressively; (2) A new competitor might’ve emerged; (3) Search volume for the keyword or query may have risen.

You can react to market changes via clear, solid visual data to support your requests for a bigger budget.

There are some cool metrics to look at. Our team has a process for reviewing and weighting metrics, as I’m sure most marketers or agencies do or will have.

Overlap Rate: How often another advertiser’s ad received an impression in the same auction you did.

Position Above Rate: How often your competitors’ ad shows above yours.

Top of Page Rate: How often you’re in one of the top four spots.

Outranking Share: How often your ad ranked higher in the auction.

What’s even better is that you can look back at an unlimited amount of time to even review entire quarters. Of course, these don’t represent 100% of impressions but at least a large enough sample.

So that’s the most significant way I know our team uses this tool. Are there other ways you guys take advantage of the data?

 

[Ryan Everson]

I hate how vendors and dealers use impression share as a barometer of how well a campaign is doing. Just like bounce rate, search impression share can easily be manipulated to whatever percentage I want.

If a vendor wants to justify more budget (and receive a bigger commission check) they can simply broaden the keywords or targeting options and show that the impression share dropped so we're losing out to our competitors and a dealership will happily approve the budget increase.

If a vendor wants to show how amazing they are (to retain a client) they can simply tighten the keywords or targeting options and show that they're achieving a 90% impression share so clearly they're doing an awesome job.

At the end of the day I would pay more attention to the cost per engagement on our site (vdp views, chat, form leads, etc) and how many store visits the campaigns generated. Too many dealers are just shoved a report with IMP, CTR, POS, IS, etc stats they don't understand or know how to evaluate and told their campaigns are doing awesome.

Now saying all of that, our corporate internal paid ads team does use search impression share and google auction insights to track our trending performance over time. We can do this because we know our campaigns are optimized well but it's not the primary metric we look at. It sounds like this is what you guys are doing Dane, so kudos to your team. For some crazy reason 20 groups are comparing impression shares now without even knowing what goes into it and that it can't be used to effectively compare different campaigns' performance.

But back to the original topic...yes it's an awesome report that most do not know even exists. Just be careful and know what goes into calculating search impression share before you use it as a primary KPI like a lot of dealers do.

 

[Dane Saville]

Thanks for the feedback!

Agreed on all of that. I didn't mean to make it sound like impression share is the only metric by which we conclude "success" or growth. This was just how we use this particular report to assist in decision-making and explaining budgeting needs and allocation of spend. There are definitely other strategies such as keyword and targeting options, consumer behavior post-click, and insights like cost per engagement.

We also dislike vendors who use tactics to increase a budget because it impacts their commission. We believe in flat-fee SEM services to avoid that conflict of interest.

You're also totally on point about dealers receiving gobbledygook that can be manipulated by a savvy marketer.

 

[Alexander Lau]

All data, in general, can be manipulated. It happens all of the time ("let's show exponential growth and they'll stay with us as a vendor, fuck the metric doesn't matter..."). Impression share is so frickin' tertiary. It's not really a KPI.

The biggest fraud in automotive paid search is the expansion of target radii.
"Oh shit, it's the end of the month and we haven't spent Dealer A's budget, let's expand their radius into areas where out-of-market buyers live." @Rick Buffkin

 

[Rick Buffkin]

@Alexander Lau, Bingo!!! I'm getting hit with display ads from dealers in PA. I live in TN!! Also to touch on @reverson point about Impression Share. Impression Share is measured when the campaign is eligible to receive impressions. Which means, if the campaign runs out of $$$$ at 1pm in the afternoon, the impression share metric stops measuring!!! So in theory your campaigns could start at 6am and be dead by 1pm and hypothetically you would show a 90% impression share for 7 hrs and thats all that will show in the reports. The remainder of the day, the missed impression share doesn't report. Alot of these vendors structure the campaigns on an Accelerated Spend meaning to spend the $$$$ as fast as possible. In certain cases this is Ok but in alot of cases, this isn't. You don't know this until you actually look at the campaign settings though. Another thing that I've seen was vendors running automated rules. I seen some vendors run an automated rule and increase keyword CPC 300% across the entire adwords account for the last 5 days of the month to simply spend out the remaining budget! There's some shady MoFo's out there!!

 

[JamieS]

@Alexander Lau, I seen some vendors run an automated rule and increase keyword CPC 300% across the entire adwords account for the last 5 days of the month to simply spend out the remaining budget! There's some shady MoFo's out there!!

I had an experience similar this once....once. Luckily I noticed it within hours, backed it up with a couple of trusted partners also seeing it, and axed the responsible party within 48 hours of them not being able to come up with a satisfactory answer.

This is probably the number one reason you have to learn GA. You must know it to an extent where you can, at a minimum, spot anomalies and act on them if necessary. Or team up with an Orbee / Clarivoy / Vistadash to help you. Never spend without having one of the two; knowledge or a partner that has your back. Ideally, have both.

 

[Dane Saville]

There are definitely SEM vendors who will take advantage of dealers who do not know the right questions to ask. You are right about the cessation of reporting upon exhausting your budget; however, I believe you can at least see at the campaign level your Search Lost Impression Share and Display Lost Impression Share due to budget. This will at least give you some ammunition to ask your vendor questions about lost share and the "why" behind any sort of accelerated spend -- it's all about asking the right question(s).

When your campaigns and ad groups are structured properly (granularly), I truly believe that impression share can give insights about how well you're bidding and budgeting for SEM, as well as understanding the overall quality of your ads and keywords (negative keywords included).

It's definitely important for dealers to be informed.

Alex and Rick, we definitely align on the idea that data can be manipulated. There can be a lot of false flags out there, both for "positive" and "negative" movements in metrics.

As far as the radius, any vendor who expands the radius to eat up a budget is a vendor I don't care to meet. It's important that the PMA sets the radius, then you can look at different areas, cities, zip codes that you know have great conversion data and potentially place extra budget there because of that data. Of course, it's also important to continue to monitor what people arriving from this channel are doing on your site -- ensuring that you're delivering an experience relevant to the search that brought them there.

Plus spending should be strategic, too. Higher impression share for intent-oriented searches; low impression share for broader searches but using automation or manually bidding up on people within your "audience" by demographic and geographic factors -- so you're not trying to spend money to win those expensive, competitive broad searches but winning buyers within those searches.

It does help, too, when your SEM vendor isn't commission-based.

 

[Rick Buffkin]

I believe you can at least see at the campaign level your Search Lost Impression Share and Display Lost Impression Share due to budget. This will at least give you some ammunition to ask your vendor questions about lost share and the "why" behind any sort of accelerated spend -- it's all about asking the right question(s).

@Dane Saville , You are right with saying, "It's all about asking the right questions". Both impression share and lost impression share are only measured when the campaign is actually running. SLIS is calculated off of Ad Rank. Ad Rank is only measured when the Ad is actually running and stops measuring when the funds are depleted for that particular day.

Personally, this is why I think using IS as a metric on an accelerated spend in the majority of cases delivers a false positive. Most vendors now a-days set up every campaign in the adwords account on an accelerated spend right from the very start. Is this a good thing? I think the smarter way is to start with a standard delivery method to get a true pulse on the IS and then adjust as needed. That's just me though.

 

[Dane Saville]

@RickBuffkin - The irony here is that I asked my paid search team the wrong question because I was not so well-versed on SLIS, so I delivered that irrelevant piece of response. You are 100% correct on depletion, SLIS, and Ad Rank -- meaning it stops measuring when the funds are depleted. I do my best to filter and deliver a message clearly, but I am imperfect!

Yes, agreed on the IS not being an accurate metric when part of an accelerated spend. We do not start on an accelerated spend from the start. It's standard delivery and adjustments as we monitor the data. We are aligned on that.

My apologies for my own breakdown in communication here.

 

[C Dorman]

Impression share, like most metrics, is an indicator. It's not a be all end all. But it is extremely useful when used correctly. I'll caution dealers to not fixate on ranking #1 or #2 all of the time. Depending on the search and the intent, paying to be #1 can be a great way to go broke in a hurry and generate a bunch of worthless clicks. The more restricted your budget, the more carefully you need to pick your fights and make every click count.