Exact match impression share doesn’t tell you what percentage of searches exactly matched the keyword.

It is
(impressions that your campaign received for searches that exactly matched your keyword)
divided by
(total estimated number of exact match impressions you were eligible to receive)

See https://support.google.com/adwords/answer/7103898

Example: Your keyword is flowers (broad match).

1) 100 searches for flowers. Your ad is shown every time. Your exact match IS is 100%.

2) 50 searches for flowers, 50 searches for roses. Your ad is shown for all 100 searches. Your exact match IS is still 100% (50/50*100).

3) 100 searches for flowers, 0 searches for roses. Your ad is shown 70 times. Your exact match IS is 70% (70/100*100).

4) 20 searches for flowers (ad shown 10 times), 100 searches for roses (shown 80 times).
Your exact match IS is 50% (10/20*100).

For exact match keywords (e.g. [flowers]) exact match IS = impression share.

Nice article – one quick question: I’m trying to analyze a large report which looks at the performance of an account over the space of 2 months. In this report I can see the KW performance for a certain day – The cost, the revenue, the ROAS etc.

I’m trying to summarize this data and show the client which ad groups and campaigns are performing well ROAS wise however since ROAS is a ratio I’m not sure how to do this. Is there another formula or method whereby ROAS can be represented as a whole number of $ amount so that it can be summarized at a campaign or ad group level?

Kind regards,

David

This is a core reason to segment search & display.

Do it 😀

I really really like your way of thinking in terms of keeping data within one sheet with multiple tabs. Your team members are lucky to have access to pre-recorded videos that form part of their training 🙂

Ouchie for the muddy reporting – but you are completely right with your statement.

I would definitely vote for having your CRO and other metrics sit within its own tab if possible.

“If you always take the top of the top (top ad, top landing page), you will be on the slow march to stronger metrics over time.” I like this approach. It makes me think of this article that sparked a new way in which I organise my campaigns: “Keywords. Treat them like employees. If they are not working, fire them” https://tenscores.com/blog/keywords/ —- It boils down to your argument of honing in on selecting the top of the top.

A bit off-topic, but kinda relates to your “don’t make me think” suggestion:
Don’t you reckon it’s just better to set custom rules/alerts or create custom scripts to eliminate you even having to import data to spreadsheets? I mean write a script TO NOTIFY the ads manager via email if keywords/ads/campaigns meet certain conditions. Even better – apply the changes if it meets certain criteria. I mean this will keep your reports even cleaner.

My next quest is to set up reporting dashboards to display on TVs within our office. Aforesaid coupled with spreadsheets = massive winnings! Klipfolio even allow you to set target goals. https://goo.gl/Cse4YB

Thanks once again for creating content and frameworks that challenge the standard 🙂

Thank you for the quick reply. The advantages of segmentation that you mentioned make sense, specially if I could come up with different ads for search and display network.

Since I currently use same ads for both search and display I prefer to combine them in a single campaign. However I noticed that when the two were combined I would get very few displays and a lot more searches. Only after I created a separate display-only campaign I started getting more display views even though the target CPA is the same for both. Do you happened to have a explanation for this?

I knew I liked you 😀
Great comment, and those are absolutely things that muddy reporting.

That said, wouldn’t you say those are separate reports for separate purposes?

i.e. if you have CRO 3rd party tools, you would simply look at CRO reporting on its own tab..

If you always take the top of the top (top ad, top landing page), you will be on the slow march to stronger metrics over time.

tldr: data is never perfect 🙁

When it comes to segmentation, it usually helps to be “as segmented as possible without impeding optimization”

So if you feel like combining the two of them is going to allow you better analysis and optimization, then you definitely should.

That said, looking at display campaigns with the mindset of “let me optimize these display campaigns” and then looking at search campaigns with the mindset of “let me optimize these search campaigns” is usually more efficient for people.

It can get messy if you constantly have to parse out a report between search & display metrics (i.e. how did this perform on search VS display). So if you segment the two at the campaign level, it saves you the extra steps of separating your data from aggregate.

Of course, there is a limit to segmentation. i.e. if you were going to have 50 campaigns for all 50 states * by 3 devices * genders -> is 300 campaigns where probably just a few would suffice. Use the rule “as segmented as possible without impeding optimization” as a guideline.

“Your conversion rate on the search network may be 3 times your conversion rate on the display network. However, your costs on the display network may be 0.25 cents per click, while your costs on the search network may be 2.20 dollars.”

Yes this is a problem if you use CPA targeting. But would it be okay to combine search & display if you use CPA targeting? This way we don’t care that search conversion rate is better or display CPC is cheaper because CPA considers both and adjusts bids accordingly.

Conditional formatting is definitely a lovely thing (see my daily report filled with it) 🙂 I like it bright and wonderful

https://uploads.disquscdn.com/images/5992fe3189f82e2c286f5ccba0c8b1b12f9749e7ea0e01db461aa965be04beff.png

https://uploads.disquscdn.com/images/53ba3ab51930739dd3fe453f63b4919b7c63e31ea7ccb0947796f78dcb87e4a6.png