I never saw this formally announced by Google; and our ecommerce clients were never notified about this change, so I just want to make sure everyone sees this update.

This is a quote from the agency newsletter:

Sunsetting Product Extensions
Product Extensions will be sunset for all enabled campaigns and will be removed from the campaign creation flow on May 20th. With increased Product Listing Ads adoption, Product Extensions have become a less relevant way to display retail products. If you have Product Extension Ad, it will continue to run as a text ad.

After PLAs were launched, Google removed the image from the extension and changed the extensions so it just showed a line of text and a price. Without the compelling image, and the additional of image filled PLAs – our CTRs on product extensions has slowly declined over the past year.

It looks like you can spend your time working on your PLA structure and making sure you are getting good PLA results as that looks like it will be the future monetization of Google Shopping. 

There was an interesting question on the forum the other day: What’s the best structure for a product listing ads (PLA) campaign? And what’s the best way to get started? As it turns out, a state of the art PLA campaign structure is very powerful, but actually not that hard to set up. All you need is the AdWords Editor, a spreadsheet, and a simple step by step process. But let’s start at the beginning…

The Default PLA Structure

Google has made sure that everyone can get started with product listing ads quickly: Create a campaign and make sure it has the product extension, add an ad group and put in the catch-all product target All products and possibly an ad. This can be done in five minutes and it’s a good way to get your first PLA campaign on the road. The resulting structure will look like this:

This setup is fairly simple and comes with some limitations. Since ads are assigned to ad groups and there’s only one ad group covering all products, this means that you can’t have different ads for different products. Bids can be applied at the product target level, but since there’s only one product target for all products the result is the same: One bid is applied to all products.

Not everyone needs different ads, but having the same bid for all products is a serious limitation. With only one bid you can’t treat your best selling products any different than your slow sellers. The bid you apply will have to work for the blended average of all products, with the bestsellers’ profits offsetting the losses from the shelf warmers.

Another drawback is that AdWords won’t show any statistics beyond your product targets. You can get some of the information from other sources, but AdWords won’t tell you anything about individual products.

A State of the Art PLA Structure

To overcome these limitations, a better structure is needed. The following setup solves all of these problems:

Here, each product has its own ad group. Inside each ad group there’s a product target that is linked to a single product using the product’s id attribute in the target’s definition.

This setup allows you to bid on individual products, which you can do at product target or the ad group level. It also gives you the most flexibility to tune your ads to your products and it lets you see stats for each individual product.

How to Build a Super-Granular PLA Campaign

To build such a granular PLA campaign, you need access to your product feed and some tools. A good approach is to use a combination of Excel (or some other spreadsheet application) and the AdWords Editor.

These are the necessary steps:

1) Create a PLA campaign and add the product extension. Note: A campaign can hold up to 20,000 ad groups. If you have more products, you will need more than one campaign.

2) Download the campaign into AdWords Editor. Add an example ad group. Then put in an example product target with id=12345. Select the target, right click, and copy it.

3) Get your product feed and open it in a spreadsheet. Make sure you have a table with all of your products.

4) Paste the product target from the editor right next to the table. This should get you a column Campaign, then a column Ad Group, then Product Target Condition 1, then Product Target Value 1 and some more. This demonstrates what you’ll need.

5) Now use simple formulas to replace the example values with real values. Leave the campaign name. As ad group, use the product’s name (formula “=B2″, for example). For Product Target Value 1 put in the ID (again, use a formula like “=A2″).

6) After you’ve done this for one product, do this for all the other products by copying the line with the formulas down next to all other products.

7) Select the columns with all the product targets and copy them. Go back to AdWords Editor and to the product targets tab. Click on the button Make multiple changes and then select Add/update multiple product targets.

Check the box top left (“My product target information below includes columns for campaign and ad group names”). Paste the columns from the spreadsheet into the field below. Then click Process and accept the changes afterwards.

8) This creates a lot of ad groups as well. Go to the ad groups tab and give those ad groups a bid.

9) Add an example product ad in the editor and use the same process (steps 4-7) to create ads for all of your ad groups as well. Using the same ad text for all products is perfectly fine, but you also have the flexibility to customize your ads.

10) That’s it – you’ve just created a state of the art PLA campaign. Don’t forget to upload everything. Make sure to keep the spreadsheet.

There is one downside to such a granular campaign: you have to keep it up to date. It’s no problem if a product is no longer available: If it’s dropped from the feed, AdWords won’t promote it. But if you have new products, you’ll have to update your targets. You can re-use the spreadsheet to do this quickly when needed.

A good addition to such a campaign is a catch-all target (All products). This way even new products are at least somewhat covered. Put a very low bid on it – much lower than all the other ones – to make sure it doesn’t interfere with your other bids.

Two Bonus Advantages of a Granular PLA Structure

In addition to full bidding control, individual ads, and individual statistics for each product, a granular PLA structure has a two more advantages. The first one is that it gives you the option to bid on the ad group level instead of the product target level. This allows you to use automated rules and AdWords Scripts, which can be applied to ad groups, but not to product targets.

The other advantage concerns the search query report. Especially with large and diverse feeds an account manager often struggles to make sense of search queries. For example, is the query “gel 3030″ relevant or should it be excluded? Having ad groups named after a product can provide some context since, in the report, the ad group’s name is right there next to the query. For example it would be easy to see that “gel 3030″ is indeed a relevant query for the product “Asics Gel 3030 Running Shoe”.

Conclusion

With the AdWords Editor, a spreadsheet and a simple step by step process it’s possible to create a state of the art PLA campaign. The process outlined here can be adapted to suit your needs. For example, you could put additional information in the ad group names, mention brands in the ads, or add custom bids depending on prices or other criteria.

So I hope this helps you build the perfect PLA campaign. If it does, or if you have questions, I’m looking forward to your feedback.

Opinions expressed in the article are those of the guest author and not necessarily Certified Knowledge. If you would like to write for Certified Knowledge, please let us know.

Search query mining is probably one of the most important tasks in managing paid search accounts. It tells us what actual search terms triggered our ads and gives us the opportunity to add or exclude those terms to improve our targeting and profitability.

Next to Google’s official explanation of the search terms report (page by Google), a lot has been written about best practices for search query mining and I can highly recommend the following articles if you want to deepen your knowledge about search query mining and match types in AdWords:

• How to Capture & Control Your PPC Keywords to Achieve a Better Account Structure by David Rodnitzky
• Advanced Search Query Mining (5 parts + Excel template) by Chad Summerhill, also see his query mining workflow over here
• The Broad Match Generator by Alan Mitchell
• Your Broad Match Keywords Are Not Converting Higher than Your Exact Match Keywords by Brad Geddes
• Match Type Keyword Trap by Craig Danuloff (written before the release of broad match modifier)

This article however, will not be about regular search query mining or match type organization but will simply show you how to find the answers to the following questions by using Excel Pivot Tables and your search query reports:

1. Are different ad groups triggering the same search query?

If you find that search queries are being triggered by different ad groups or even campaigns, you are competing against yourself. Besides, the same search query sees different ads and possibly even different landing pages, so you’ve lost some control of what you’re showing to whom.

By knowing where this happens, you can fix this by choosing the ad group where the search query belongs and excluding the query in all other ad groups.

2. How are the close variant queries performing in your campaigns?

In April 2012 Google announced a new matching behavior for phrase and exact match keywords, which now also include plurals, misspellings and other close variants, unless you opt-out.

Whether or not this ‘near match’ is actually working for you often depends per campaign. If it’s giving you more volume within your efficiency target, you should leave this option on. And if it’s not, then you should obviously disable it.

By knowing how this keyword matching option is performing in each of your search campaigns, you can selectively opt-out based upon your own data.

Finding different ad groups that target the same search query

To find out if multiple ad groups are triggering the exact same search query, follow these steps:

1. Go to the Keywords tab on the account level and select a significant date range. The past 3 months usually works well, but feel free to use a shorter or longer time range depending on the size of your account.

2. Get the search query report by clicking on the “Keyword details” button and selecting “All”.

3. You only need the following columns, make sure you have these by clicking on the Columns button and selecting “Customize columns” (and feel free to remove the other columns):

4. If you have separate campaigns that target different devices, geographic locations, time of day or day of week, then create a filter to make sure that you’re only viewing queries from campaigns that share the same targeting settings.

You can do this by clicking on the Filter button and selecting “Create filter”. Now you can manually add the campaigns with shared targeting settings one by one, or filter them by campaign name if you’ve clearly labeled your campaigns.

For example, you should exclude campaigns targeting mobile devices when you’re analyzing the queries of your computer (and tablet) campaigns:

5. Download this report in Excel .csv format and open it with Excel

6. Delete the first and last row of this report

7. Insert a Pivot Table based on this report

8. Select the Search term, Campaign and Ad group fields (in this order) in the Pivot Table Field List, so these appear in the Row Labels of your Pivot Table.

9. Drag the Impressions, Clicks and Conversions fields (in this order) to the Values part of your Pivot Table.

10. Click on each of the values and change the “Value Field Settings” to “Sum” (instead of “Count”)

11. Insert a calculated field: CTR (we need Excel to calculate this, inserting the AdWords CTR data directly will not give us the correct numbers):

Excel 2007:

Excel 2010:

12. Click on the arrow in the Row Labels field of your Pivot Table and select “More Sort Options”

13. Sort “Descending (Z to A) by”: “Sum of Impressions” to get the search queries with the most impressions on top.

14. Now look for search queries that occurred in different ad groups:

15. Now it’s up to you to decide in which ad group the particular search query belongs and to exclude it as an exact negative in all other ad groups (or campaigns).

Finding out how near phrase and exact is performing per campaign

Follow the same first 6 steps as in the previous part (skipping step 4). Then continue with the following steps:

1. We want to consolidate the data of the following match types: “exact (close variant)” and “phrase (close variant)”, as we can only enable or disable near matching for both match types at once.

One way of doing this is by putting a filter on the first row and selecting just both close variant match types:

Now you can easily replace both match type names by “close variants” by typing this into the first cell and double clicking the fill handle (the skinny plus sign that appears in the bottom right corner of a cell).

2. Insert a Pivot Table based on this report

3. Drag your Campaign and Match type fields (in this order) into the Row Labels of the Pivot Table

4. Drag the fields of your choice into the Values field, as long as they’re ‘raw’ numbers (so no calculated metrics from AdWords like CTR, CPC, CPA and Conv. Rate), for example: Impressions, Clicks, Cost and Conversions

5. Click on each of the values and change the “Value Field Settings” to “Sum” (instead of “Count”)

6. Insert a calculated field: CPA (we need Excel to calculate this, inserting the AdWords CPA data directly will not give us the correct numbers):

Excel 2007:

Excel 2010:

7. Now your Pivot Table shows you performance metrics for each match type within each campaign.

8. In the example above we can see that in Campaign 1 the close variant match type has a significantly higher CPA ($101) than the campaign average ($74) and especially than exact ($67) and phrase ($71).

In Campaign 2 however, we see the close variant performing better than the campaign average and even better than the phrase match type.

9. It’s up to you which difference in performance you want to tolerate before opting out of is this keyword matching option, but I wouldn’t advise to opt out by default as it often brings in additional traffic for reasonable CPA’s.

Some have found that singular terms convert better than plurals (based upon last-click attribution) and I also see this happening in our accounts. This could be part of the explanation of the difference in performance of close variants. In those cases you would want a different (bid) strategy for plural and singular keywords, which means you’ll have to opt-out of this keyword matching option.

Conclusion

I hope you will gain new and valuable insights by using your search query reports and Pivot Tables as described in this article.
If anyone has any other creative uses of AdWords reports and Pivot Tables, I would love to hear them!

 
Opinions expressed in the article are those of the guest author and not necessarily Certified Knowledge. If you would like to write for Certified Knowledge, please let us know.
 

A Search by Any Other Name

Have you ever looked through your AdWords search query reports and seen terms that are utterly baffling in their structure with lots of impressions? For instance, things like (by the way, these are all real queries I’ve seen):

• for the home bar stools for the home
• social work more:label_social_20work more:gradschool
• property deeds more:label_property_20deeds more:homebuying
• women’s fashion clothing shoes dresses shirts

So who are the half-million like-minded oddballs who searched for “women’s fashion clothing shoes dresses shirts”? The answer is that they don’t exist; they didn’t search for that term. These search partner ads were served on what is technically a content site. For certain ecommerce sites and other sites with a taxonomy-based site structure, Google uses the category string of the page as a keyword and serves “search partner” results for views that were not literally triggered by a search.

Don’t Freak Out

When I first became aware of this a couple of years ago, my initial reaction was outrage (“How dare they frame these content results as search results!?”). However, after giving it some thought I realized that there are only two ways that people can end up at these deep category pages:

1. Through clicking an organic result after making a highly relevant search
2. By a very pointed category drill-down on a site

Either one of these cases is very intent-driven and much more analogous to a search than simply “happening” upon some content while aimlessly browsing. Additionally, though the results from these types of clicks are not always stellar, the conversion numbers these typically bring are in line with what you would see from any other “search partner” site. So, I came around to feeling it was pretty fair to call these “searches.”

What to Do When You See These Terms

The atypical queries in your search term reporting mentioned earlier are one way to find these scenarios within your account. However, they are just the “tells.” In these easily identified cases, optimizing is straightforward:

1. If the query is producing undesirable results, simply use a negative keyword to block it. Because these are “searches,” not content, negative matches are clean and will prevent ads from showing.

2. If the results are good, isolate the query as an exact match and bid to your target metrics. Something to be aware of when using this strategy is that often there are a limited amount of ad slots for these, sometimes only one slot. Because of this, I usually add a tag to these keywords and call them “all or nothings.” You’ll want to keep a close eye on them if you are relying on this traffic, as slight variations in your bid or the competition can call these to fall off the limited ad space and you can lose huge amounts of impressions over a couple-cent bid change if you fall out of the results.

More problematic is that often the category name that Google uses as the trigger keyword is a regular keyword; in fact, it is often a head term where the impact can get easily lost in all of the impressions.

The above is an example of this type of ads on Macys.com for the “women’s shoes” category. Take note that these ads have some features that true content ads would not, including sitelink extensions and bolding of the keyword in the ads. In this case, the keyword is “women’s shoes.” You will not be able to fully map your results on these head terms to this particular phenomenon; instead, you have to rely on search partner reporting as a blended metric inclusive of “regular” search partner results. That said, you can still find cases where this is likely happening.

1. First, you need to make sure that you have tight mapping of your keywords and ad groups (at PPC Associates, we generally use single-keyword ad groups to do this).

2. Then you can segment your ad group level report by network inclusive of search partner data. When you download, add an additional segment for reporting by week.

3. By looking at the search partner data week over week, paying attention to your overall impressions and your positions, you can pick up where this is probably happening.

In this example data, note that when my average position slipped from 2.17 to 5.38 I lost 65,000 impressions a week. This is a pattern not typical of “regular” searches and indicates that this head term “women’s dresses” is highly influenced by these category-type search partner impressions.

These head-term cases are much harder to optimize for. It is one more reason to wish Google would allow search-partner-only campaign targeting – or, even better, reporting and inclusive and exclusive targeting by specific search partner site. In lieu of those tools, you have only a couple of choices:

1. First, isolate the head terms that are influenced by this in their own campaigns.

2. If the results are undesirable, eliminate search partner traffic entirely on this campaign. The down side here is that there may be some good search partner traffic that you are getting rid of.

3. If your conversion metrics are good on this search partner traffic, you need to pay attention to maintaining your positions so you don’t lose your impression share on these keywords. Consider a bidding strategy that not only looks at your ROAS or CPA goals, but also factors in staying above a position threshold where you don’t fall off of the available ad slots on the search partner network.

Stay Vigilant in Q4

Though this phenomenon is not isolated to ecommerce sites, both on the advertiser and site partner side, it is definitely most prevalent for ecommerce sites, so I highly suggest taking a deep dive to look for how this is influencing your campaigns if you are running PPC for an ecommerce site. Additionally, I would expect this to become even more prevalent as we enter Q3 and Q4. I have seen sites such as target.com, jcpenney.com, sears.com, and amazon.com all serving these category-as-search-string ads, though as I write this, none of them seems to be currently using this practice. My speculation is that as we get closer to Q4 and the available revenue heats up, we’ll see these ads pop back up on many sites that are not currently using them.

 
Opinions expressed in the article are those of the guest author and not necessarily Certified Knowledge. If you would like to write for Certified Knowledge, please let us know.
 

Bidding is a cornerstone of Paid Search management and the dynamics of the ever-evolving auction marketplace auction marketplace, combined with the multitude of bidding strategies, make the search marketing professional a key element of a well spent media budget. Bid optimization is also an intricate and dynamic assignment with many considerations which increase with the size and complexity of your accounts.

I’m assuming you have conversion tracking set up through the Adwords interface and ideally some sort of Revenue data as well. The first should be considered mandatory, the second a nice-to-have. If you don’t have an Adwords goal set up, “you’re doing it wrong” and should likely start there first.

The Three Bidding Options

If we ignore neglecting your accounts entirely (not recommend), there are three basic options when bidding your keywords:

Manual Bidding

• Keyword by keyword bidding in the interface:

• The simplest form of bid management. The most time consuming, and often the least effective in that in very large accounts you may rarely get to the longer-tail keywords, or in very dynamic accounts, Display Network or other specialized campaigns may be neglected.

• Bulk bidding in the Adwords Editor with filters:

• Can be very effective for bulk bidding actions, although I prefer the interface. Note that this can actually be a good intermediary step to help determine your automated rules. More on both of those later.

• Export to Excel:

• Export, perform calculations, upload in bulk to the Editor. Also more scalable, although likely more complicated than necessary since the same result can be accomplished in the Editor, although it makes segmentation of keywords more visual and pragmatic. It may also be necessary if you are tying in revenue data from a source other than Adwords.

Automated Bidding

• Third-Party Tools:

• While I don’t have an inherent bias against third-party bid automation, in my personal experience I have yet to find a solution that I felt completely comfortable with. But frankly, I don’t employ them often enough to make a strong case either way and this isn’t the primary purpose of this post.

• Google’s Non-Goal Focused Automated Bidding:

• Google manages the bids and attempts to acquire the most clicks for our budget. That should sound scary to anyone who grew up in the post-traditional advertising world.

• Google’s Conversion Optimizer:

• Used correctly, I have found Google’s conversion optimizer to be a fantastic method for managing portions of your accounts. My preference is using Conversion Optimizer after considerable pruning and granular ad group segmentation, on very high conversion volume ad groups whose search volume does not substantially fluctuate. It is also useful when there are a number of factors at play in the campaign such as multiple devices, locations, languages, etc.

Manual Bidding with Automated Rules

The subject of this post and the type of bidding that I am moving more and more towards each passing month. Let’s discuss why.

Why Manual Bidding?

There is nothing inherently wrong with automated bidding; the problem is that bidding does not exist in a vacuum. In a perfectly tailored and segmented account, you could determine a keyword’s performance, calculate its revenue per click and bid based on the desired margin.

However, there are so many reasons why a keyword is performing a certain way that until you have done ample creative testing, segmented your campaigns by any relevant dimensions such as device, network, location, etc. and scoured search query reports for ad group segmentation and negative keyword possibilities, you could be making decisions based on blended data. As Analytics savant Avinash Kaushik says, “all data in the aggregate is useless,” and that’s exactly what you’re looking at when many different factors roll up under an ‘individual’ keyword.

Even when you’ve done all the things mentioned above, you still have the problem of attribution, especially for higher funnel keywords which may not be getting the proper credit for their influence on conversions down the funnel. Finally, automated bidding often doesn’t work very well for keywords with sparse data, which will likely be 60-80% of your account.

Thus we are left with manual bidding for much of our account, even when using Conversion Optimizer for some campaigns. This is where automated rules can become your best friend, a huge time saver and an efficiency driver to garner both increased performance and more time at the beach.

When Your Account is Ready for Automation

After bidding a particular account for a while, I start to zero in on certain factors by which I make bidding decisions and somewhat standardized responses based on the data. When you begin to realize, for example, that you use a seven day date range on higher volume keywords and bid down around 10% when the CPA is $4-6, you’re ready to automate some of your efforts.

In a nutshell, automated rules are best used in place of repeated basic actions in response to a certain threshold of data.

Set Volume Thresholds

High-, medium- and low-volume keywords require different strategies to be bid properly. To define these categories a bit more, though they are loose, flexible and personal designations, a high volume keyword is one that acquires statistically significant impression, click and conversion volume on a daily to weekly basis that is representative such that the keyword can be bid entirely based on this recent data – wild seasonal fluctuations or temporary site issues notwithstanding.

Medium-volume keywords are those that acquire enough volume and conversion data to be bid on two to four weeks of recent data, and low-volume or long-tail keywords are those that require multiple months of data to accumulate effective bidding data. I’ll leave cross-ad group data aggregation, statistical models, and accounting for year-over-year seasonal changes to other posts.

It is important that you take these loose buckets into account when setting bidding rules and I do not recommend creating automated rules without volume thresholds.

Factors for Bucketing Your Keywords

I have found the below to be the most common filters by which I set my automated rules.

• Date Ranges:
  • Lower volume keywords require longer date ranges.
• CPA or ROI Goals:
  • Your overarching goals for the account are likely what most of your rules will be based on.
• Conversion Volume:
  • Keywords with low conversions (1-2 in the given date range) should receive lighter treatment when using CPA or ROI goals (ie. If you’re increasing keywords with a CPA of $3 or better, you might not increase as aggressively on a keyword with $3 spent on 1 conversion, versus a keyword with 3 conversions at $9 of spend).
• Impressions:
  • Bidding activity is often biased towards bidding down. Most keywords perform poorly during one time or another, and this will often lead to decreased bids. As such, some keywords will only be bid down until they lose most of their volume. Light automated bid increases on low impression keywords can help restore lost volume for keywords.
• Quality Score:
  • You may want to take Quality Score into account when choosing keywords to bid up, especially on lower volume of clicks or impressions. Higher quality score keywords obviously have lower CPCs, as well as more likelihood of showing in top positions, showing up at all, and/or displaying various ad extensions.

Categories and Examples of Automated Bidding Rules

Let’s get into the actionable meat of this post and discuss some actual rules you could consider using.

High Volume Keywords

Run: Weekly
Date Range: Last 7 Days
Volume Thresholds: Typically greater than 10 clicks, often higher.
Performance Thresholds: If you’re target CPA is $10, you might bid keywords over ~$12
Action: Bid down 10-20%

Sister Rule: For keywords below the target CPA (say below $8 for this example), bid up 10%.

Notes: As noted above, I like to give some leeway above the allowable CPA. I don’t want to be too aggressive on a keyword that has a bad week and is within a small percentage of an allowable CPA with just one week of data. I also like to bid up or down just 10-15% for the standard rules; you can save more drastic changes for more drastic losses. Often I will also require a minimum of 1-2 conversions, and create another rule for those with 0-1 conversions that is tied to spend for that date range rather than CPA or ROI.

Mid/Low Volume Keywords

Run: Once per month
Date Range: Last 30 Days
Volume Thresholds: Some click threshold, at least over 10 clicks typically
Performance Thresholds: Similar to above
Action: Bid down or up based on performance, ~5-15%

Notes: I like to be less aggressive on this bucket as it will often be a large number of keywords and so have a noticeable effect on the account anyway.

Catch the Bleeders

Run: Daily
Date Range: Yesterday
Volume Thresholds: Typically a cost threshold which depends on your average spend and CPA/ROI goals
Performance Thresholds: No conversions with high spend or enormously high CPA
Action: Bid down 20-50% or pause so you can review

Notes: I’ll use a rule like this, which rarely runs, to catch a runaway keyword which could cost hundreds or thousands of dollars if left unchecked, often as a result of launching a new keyword, or accidently entering a bid of $5 when you meant $0.50 (happens to the best of us from time to time!).

Mid Volume or Very Low Volume

Unfortunately, Adwords currently only allows you to run on a weekly or monthly basis, which leaves some gaps for mid-range keywords. And for very long-tail keywords, although you can bid on 60 or 90 days of data, only being able to run these on a monthly basis can create some overlapping bid changes on the same data as used previously. Therefore you may choose to take advantage of the ability to run rules only one time, on command, for these buckets and do so within a given month or every 2-3 months for certain buckets of keywords.

Be Careful of…

Rules Without Bid Caps

I highly recommend setting maximum and minimum bids for your rules based on the average range of CPCs for your account. If a bid of over $3 seems outlandish for your account, cap the bid there; you can always increase it manually if it is warranted.

First Page & Top of Page Bid Estimates

I recently inherited an account with an automated rule that raised every active keyword to the first page bid estimate every single day. Rules like this are why Google set up automated bidding rules, as can be inferred by how prominently they are advertised, in order to drive up the average revenue per click in their auction.

These bid estimates are very ballpark, rounded, and often very inaccurate in practicality. Imagine if multiple advertisers in your industry used rules like this and all your competitors responded to these bid increases; suddenly your bids are being increased in perpetuity over time. You know who wins here? Google.

Overlap

You should try to ensure that your keywords are not touched by multiple rules or overlapping date ranges; with many rules, overlap will be somewhat inevitable given the limitations, but as long as it is minimal you should be fine.

I hope this helps you strategize ways that you might use automated bidding rules to both ease your burden as a search marketer and achieve better campaign performance for your business or your clients’ businesses. This is a new an evolving methodology for many search marketers. Feel free to offer more ideas for rules you’ve used in the comments below!

 
Opinions expressed in the article are those of the guest author and not necessarily Certified Knowledge. If you would like to write for Certified Knowledge, please let us know.
 

The main factors for Quality Score haven’t changed in a few years. The application of those factors have changed over time with the two most notable changes being:

• Landing pages changes in October 2011
• User interface refresh in April 2012

Google recently updated the Quality Score algo to take into account the ad’s performance on a device:

• Desktops/laptops
• Tablets
• Mobile phones

So, if you have a campaign targeted to all devices, you’re not really seeing your Quality Score as it relates to each device.

I’m a huge fan of segmenting campaigns by devices, assuming you’re not making so many campaigns that you thin your budget too much and spend more time in budget management than in AdWords optimization.

If you can’t split out your campaigns by device; then you can at least segment CTRs by devices by using the segment feature.

Once you segment by device, you can quickly see at a campaign or ad group level if you have large differences in CTR (and for optimization, conversion rates and CPAs) by device:

If you start to see differences; then you can even segment the ads themselves by device type:

In many ways, this change should only force you to make better decisions with your ads and landing pages by device type. User behavior is different on a phone versus a desktop; and the landing pages need to be different as well (unless you have a very nice responsive web design) based upon that behavior.

What’s really being obfuscated here is that you might have an ‘average’ landing page; but its really above average on a laptop and below average on a phone; but you just see the combined results.

If are advertising on multiple device types and you haven’t segmented your campaigns by device yet; then it might be time to put that change on the agenda.

Note for Premium Members: The Quality Score videos are already being refreshed and this new data will be in the video updates very soon.

Many of us obsess over Quality Score – sometimes too much. There are times when you should completely ignore Quality Score and just make sure that you are providing a good user experience that culminates into new customers.

This is bound to be a little bit controversial, so please feel free to disagree and leave comments when you think I’m wrong. This is based upon looking at way to many accounts where Quality Scores just don’t make sense.

There are three main times when you want to ignore Quality Score:

• Google isn’t calculating it correctly (bugs)
• You don’t have enough data
• Your industry is just messy (mostly finance, pharma, and legal)

With each of these scenarios you shouldn’t give up before you try. You might try a few variations of ads, split out the keywords into their own ad group with their own ad and landing page, etc. And after a few rounds of tests, if you can’t make the Quality Scores, then you should walk away.

I’ll walk through the main issues I see where a keyword has a low Quality Score and you should just walk away and not obsess over it after you have done some testing.

Word Variations

These scenarios are probably the most maddening. Certain words don’t have different CPCs, CTRs, or conversion rates by plural, singular, or various stems. Yet, the quality scores are completely different for each variation.

If you see any of these happening in your account – you are not alone.

Plurals & Singulars:

• Your singular (or plural) is a 7-10
• The reverse, your plural (or singular) is a 3-4
• For example: The word ‘accident attorney’ is a 7-10; the word ‘accident attorneys’ is a 4

Stemming:

• The word is a 7-10
• Any stem is a 3-4
• For examine: Gardening tools is a 10 and garden tools is a 3-4.

Reverse words:

• Accident attorney is a 7-10
• Attorney accident is a 3-4

With these words, if your CTRs, CPCs, and average position are similar for both variations then after a few tests walk away. I can’t count the number of wasted hours I’ve seen on these tests.

You Don’t Have Enough Impressions

This is one of the biggest mistakes I see. Someone adds 100 new keywords, immediately gets 3s and then starts to worry about it.

Or, someone adds 1000 keywords, waits 24 hours, sees lots of 4s and starts testing when no keyword has more than 30 impressions.

Your keywords don’t start with your Quality Score. They start with an average that is tempered by your overall quality score.

Don’t even look at Quality Score until you have 500 impressions. If it’s a higher volume term, then wait until you have 1000-2000 at a minimum. The lone exception is when you see below average landing page experience. In that case, take a look at your organization and landing pages. Although, landing pages are hard to diagnose right now. You can have  10 with a below average landing page:

However, if you add lots of words and most of them are 7s and everything is average; and just a few are 3/4s and they have below average landing page experiences; then take a closer look into the landing pages.

It’s Just Your Industry

Some industries have had such a plethora of bad traffic bought within them that no one can get a good quality score. This can be maddening as you think you should have 7s; but no matter what you do, you can’t  break 3-4. Some accounts are in great shape with an average quality score of 3.5. This is not common; but here’s how to diagnose it.

What you will often see is that you have high positions; but low quality scores.

If you see this scenario, then take a look at the ‘top vs other’ segmentation for those keywords:

Google does not put ‘bad ads’ above the organic results. if you see your ad above the organics, and you have quality score 3 and 4 words – its most likely your industry. This is very common in these industries:

• Financial
• Pharmaceutical
• Legal
• Industries with average CPCs above $35-$50

With these words, there is one big issue:

• At a QS 3, your ads can be above the organics and show all the time
• At a QS 2, your ads are rarely shown

There is a huge difference between a 2 and a 3 Quality Score. Therefore, you will want to get to 3s. A 4 is good (consider it a 7), and a 5 is great (consider it a 10); and rarely does anyone have an actual 7 to 10.

Words with Other Meanings

Not everything means what you think it does. Many words have multiple meanings; and this can cause low Quality Scores due to low CTRs and you usually don’t want to fix them as doing so means you get lots of unqualified visitors and your conversion rate suffers.

It is better to have a low quality score, yet profitable word, than a high quality score that’s just costing you lots of money.

Demographic Query Data

Often words are not searched (and therefore clicked on) by your target market. I still see people wondering why they can’t get a good CTR on a word like ‘bleach’ in the cleaning industry. That’s because the majority of people searching for ‘bleach’ are actually looking for Japanese Anime and laundry supplies.

In a case like this; you’ll never get a great CTR. You can do lots of negatives; but be prepared to have some low CTR or to use such generic ads your conversion rates suffer. Always know who is searching on your ads – you don’t always want a high QS if you lose money by having one.

It’s Someone’s Brand

You see Quality Score at the keyword level. Quality Score takes into account geographic & ad CTRs. If you are advertising on a word that is a brand; you might have a low CTR in some regions, and a great one in other regions. You will just see the average. In this case, you often end up with a low Quality Score if that brand has enough of the search queries and clicks in the majority of other regions.

It’s a Common URL

There are a surprising number of searches for URLs or common company names. Not everyone directly goes to a website (often called direct navigation). It’s pretty common to see people search for a site and then click on the link to go to the site.

While these examples are a bit extreme (why are you on Google searching for Google anyway?); if your keyword is a popular URL; then you will often end up with a low quality score. Much like the brand names above, it is due to a low CTR. However, unless you’re ready to spend some money branding that term, accept the low Quality Scores.

Conclusion

I must reiterate – don’t just look to see if you’re in one of these examples and there never try. Sometimes you can get good quality scores in these scenarios. However, if you have tried and failed, and yet your CTR, conversion rate, and CPCs are the same (especially in the instance of word variations); then do your self a favor and go do some testing in another ad group and save yourself time.

Assuming you’re not in one of the industries that has real issues, as long as 70% of your impressions are in the Quality Score 7 and higher region, then you’re in good shape and don’t let a few bugs ruin your day. A healthy account can easily absorb the traffic impact from a few low Quality Score keywords.

If you need help determining how healthy your account is, please  see this video: How to Identify Google AdWords Quality Score Problems.

Quality Scores are important. However, there comes a point when your time is better spent elsewhere. These overall guidelines should help you determine when you should walk away from Quality Score testing to spend your time on something more productive.

Google’s near match or close variant match is now live. Some accounts already have it; and others should be seeing it soon.

If you missed the announcement, here’s the gist:

• The close variant option is a campaign setting
• The setting is turned on by default
• Your phrase match or exact match terms can be shown for ‘close variations’ of keywords

This means exact match could be exact match or close variant exact match. Previously, I wrote some initial impressions from a conceptual standpoint; so I’m going to skip that for now.

These new match types are designated in the search query report as:

• exact (close variant)
• phrase (close variant)

What is Being Analyzed

I wanted to do some initial analysis to see how this match type variation is performing for a few accounts. As this match type has not been live very long yet; don’t take this as gospel. I only examined $680k in spend (just the last few days of data in a few larger accounts) in creating this analysis.

In addition, I’m comparing the traditional match type to the close variant match type for CPA, conversion rate, and CPC.

I’m not examining the incremental conversions yet. As for that analysis, I need to examine when an exact or phrase match variation also existed in the account; so the close variant wasn’t actually adding additional volume. There is a lot of de-duplication that needs to occur so I want a lot more data to work with before I undertake that huge analysis.

In addition, session-based broad match is not new. But since this was an analysis of a variant match versus the traditional match; I thought it would be useful to include it since its already part of the search query report.

The Analysis

I wanted to make this as easy to read as possible since there are lots of stats. So, what I did was examine the delta between the match type and its variant for:

• CPA
• Conversion rate
• CPC

Then I wanted it to be easy to see what was best. So the deltas are color coded:

• Green Arrow: The variant is at least 10% better than the traditional match.
• For CPC, this means its 10% lower
• For CPA, this means its 10% lower
• For Conversion rate, this means its at least 10% higher
• Yellow bar: The variant is between 10% lower and 10% higher. I’m just calling this a push for now
• Red Arrow: The variant is at least 10% worse than the traditional match
• For CPC, this means its 10% higher
• For CPA, this means its 10% higher
• For Conversion rate, this means its at least 10% lower

The Data

As these accounts have different CPCs & CPAs; I didn’t want to combine the data together, so first I’ll show all the data individually, then sum up the results. You can click on any of these images to see them in a larger size.

As this match variation has only been live for a few days, these stats are only from the past few days.

The first account is for a subscription membership site:

The second account is for a lead generation site:

The third account is for a toolbar installation:

The fourth account is for product samples:

And this is for one of those ridiculously expensive industries:

The Results

In total, this is what the results look like (and remember, its only 5 tests – so its initial data; but not conclusive):

It appears at just a quick glance that exact match is better than the variation of exact match. While the CPC is lower for the variations of exact; the CPA as conversion rates are better for exact match.

However, it looks like phrase match variation might be a push against phrase match. On the CPA side, phrase match variation was the same or better. It appears that the reason the CPA is so similar is because the phrase match CPCs are lower; so even while the conversion rates are better for phrase match, the lower CPCs make the CPA a push. As near match rolls out to everyone, I would anticipate the CPCs will go up and therefore make phrase match better than the variation in the long run, but that has yet to be seen.

I was surprised on the conversion rates of the session based broad match. It had been a long time since I’ve done that analysis. However, as you can’t control if this is on or off; there really isn’t a point in doing a huge analysis other than to present non-actionable data.

Analysis Needs Incremental Conversions

This is an initial analysis; but its not conclusive for a few reasons:

• This is just a few days of data. More data across more account types will give more accurate results
• Not all accounts are being shown for variations yet; so I’m assuming CPCs will rise as competition for near match goes up
• This analysis doesn’t examine incremental conversions

My real question is: If the account has lots of exact and phrase match keywords, are these new conversions or just reassigned conversions?

In examining a few ad groups, I’m seeing a bit of both; but to what extent I don’t know as I really want a solid month of data when every account who will be in the program (assuming you didn’t opt out) has this feature enabled to see the long term results on incremental conversions.

If you want to learn more about how this match type works and learn how to opt out, please take a look at this article: Google Launches Near Match – Should You Use It?

Conclusion For Now

The near match variation is better than I expected; especially for phrase match – so this is not a setting to blindly turn on or off.

If you have a set budget, and the near match is taking budget away from your other keywords at higher CPAs / worse conversion rates – then I’d turn it off.

If you have a set budget, and the near match is a similar CPA; then you might just play a waiting game at the moment.

If you have budget left that you’re not spending; then I’d wait until you have a little bit more data to see the CPA of incremental conversions before making any final decisions.

Search advertising, specifically Google, Yahoo!, and Bing, is successful because it allows marketers to reach consumers who have expressed direct interest in their product. This extends to targeting options on the Google Display Network as well, since, by and large, GDN is one of the few major ad-serving platforms that have turnkey semantic targeting available.

For the last year-and-a-half or so, however, Google has identified two big problems with ad targeting options on GDN. For one, they’ve realized they’ve got a lot of ‘remnant’ inventory, which they had been selling for extremely low CPMs (think video gaming websites, myspace, dictionary.com, etc.). The solution here was remarketing, and audience-level buying, which allowed high-CPM bidders to serve ads on traditionally low-CPM placements.

The latter of these two, audience-level buying, also helped address the second problem, which was satisfying the needs of more traditional display advertisers. Think along the lines of a Groupon, OneKingsLane, or Netflix – companies that didn’t have ‘hand raisers’ like the traditional direct-response company. These companies are new mousetraps, and that’s big business, so Google answered with less granular, more scalable targeting options like the aforementioned audience buying, as well as topics and demographic ad buys. These media buying options also spoke to big-brand advertisers, who previously were burdened by the limited scale of GDN when it only had semantic targeting options.

So, to completely over-simplify, direct response companies love semantic targeting, and ‘new mousetrap’ and brand advertising companies prefer the scale of audience- and topic-level buying, especially when give the opportunity to layer on demographic info (something on which Google is working especially hard).

The problem lies when Google decides to blur the lines between two advertising approaches.

On the surface, it appears that Google has done a great job separating the targeting methods for these two type of advertisers, but over time, I’ve seen that, in fact, there are several instances when an advertiser thinks he/she is doing a semantic buy but is in fact relegated to a purchase not based on the content of the placement. I’ve found three main situations where this behavior occurs.

Keyword Parsing

If you’ve ever bought a keywords like ‘Atlanta medical schools’ in the GDN and found that your placement report had domains and urls that only had content related to ‘Atlanta’ and nothing to do ‘schools,’ let alone ‘medical schools,’ you’ve been a victim of keyword parsing in GDN. This is when Google determines there isn’t enough volume in a particular ad group, and does the advertiser a ‘favor’ by deciding to ‘cut’ keywords at the token level and split them into multiple semantic categories. So, in the above example, you could be matched to pages about ‘Atlanta,’ ‘Medical,’ or ‘Schools.’ If you speak to any display specialist at Google, he/she will note this as a common occurrence, though I’ve never found a record of it any of Google’s literature.

In search, you’ll see this occur in a less explicit way with broad match keywords.

How to Avoid Keyword Parsing

In the search network, you can avoid keyword parsing by using more restrictive match types, such as modified broad match. Additionally, consider negative exact match usage. For example, if you have a 3-token keyword being matched to irrelevant 1- or 2-token queries, consider making the shorter version exact match negative.

The bigger issue is in GDN. To limit keyword parsing in GDN, avoid using geo-modified keywords (or other modifiers that aren’t relevant as stand-alone tokens). Also, create ad groups with a clearer semantic theme by adding in a few more keywords. Lastly, and most importantly, make use of negative topics, categories, and domains. Typically, a topic that could very easily be made negative by most direct response advertisers is the ‘games’ topic. However, avoid audience-level negatives, as excluding an audience like ‘games’ will exclude a lot of great potential customers – because, let’s face it, we’re probably all a part of the ‘games’ audience!

Automatic Targeting Option Extensions

When advertisers create keyword-targeted campaigns in GDN, I’d imagine it’s typically because those advertisers only want to serve ads on content pages related to their keyword’s semantic theme. Unfortunately, Google doesn’t quite agree.  Here’s a note from one of my better Google reps:

All Google Display Network campaigns which use contextual targeting are
given the opportunity to appear on as wide a range of relevant pages in
the network as possible. In addition to the factors which we more commonly
use to match these ads to the page (e.g. page content, language, link
structure, page structure, URL), we take into account additional
contextual signals, such as the content of a page which a user recently
browsed. The aggregate of all of these signals, and the performance and
bid(s) of each of the ads, ultimately determine which ads appear to a user
for a given impression.

What this means is that Google can extend semantic targeting to ‘audience’-level targeting at will. I don’t understand this at all; if advertisers wanted to do audience-level buys, they surely would. The result of what I’ve coined as ‘audience-targeting extensions’ are placements on domains that have absolutely nothing to do with the keywords in the ad group! I’ve found this occur more commonly, for whatever reason, with image ad formats. Additionally, this is much more of a problem when you’re in a vertical with high CPM or CPC bids. If you’re in one of these verticals, this is something you should really pay attention to. I’ve seen up to 80% of spend coming as a result of an audience-level extension (I’ve seen this twice in fact, both times in the legal vertical. I also see this in EDU verticals to a lesser extent). The results of the campaigns were, predictably, extremely poor.

In search, automatic audience extension exists as ‘session-based’ ad serving. The story is the same here: the advertiser was automatically extended to show ads to people based on their prior behavior, not based on what their current interest has been expressed as.

How to Avoid Automatic Targeting Option Extensions

This one is tough in search, though it’s a pretty ‘small’ problem in most instances. Unfortunately, there is no real lever I’m aware of to control this, and I’ve been relegated to waiting for an ‘opt out’ session-based ad option from Google.

In GDN, the options are much the same as with keyword parsing. The main one, again, being topic-level exclusion.

Semantic Reaches

Semantic reaches can also be coined as ‘broad match gone wild’ – essentially, it’s when a relevant keyword gets matched to an irrelevant query. In fact, semantic reaches were once such a problem that most advertisers avoided broad match completely. With the advent of modified broad match, however, semantic reaches aren’t in the spotlight nearly as much.

As Google’s search algo understands semantics better, another problem has arisen. Typically, advertisers bid very high on their brand keywords to ensure position 1.0. Additionally, they are highly concerned with capturing 100% of their brand traffic, and hence run broad match. In this scenario, both adCenter and AdWords will match brand terms to relevant non-brand queries, the result being incredibly high CPCs for queries that weren’t intended to be bid that high. I’ve shown this behavior to some of my clients who are new to PPC, and they are shocked! Google’s algo is really advanced in this area, so if you’re ‘Netflix,’ Google will likely match your brand to things like ‘movie rentals’ and other similar queries. If advertisers aren’t checking their search query report, this problem often goes unnoticed, because the true brand queries cover up the poor performance of the non-brand queries, and the campaign and keyword level CPA (or ROAS) appears just fine.

How to Avoid Semantic Reaches

Avoiding semantic reaches is relatively easy. For brand queries, I’d advise bidding high, but still capping them at reasonable max CPCs. So, say you’re paying $0.30 for a brand click (position 1). In that scenario, bidding $1 would provide the same advantages of bidding $100, so I’d choose the former. Additionally, migrating away from broad match to modified broad match is a very safe approach. In this scenario, you’d likely capture 100% of the traffic you intend to capture, without the risk of capturing some traffic you didn’t intend to capture.

Conclusion

I don’t think there’s much dispute for direct-response advertisers that GDN is the best available display network, in general, for creating on-click (as opposed to view-through) sales. However, as Google starts to cater more and more towards brand advertisers, my worry regards the attempts to blend the two approaches. As an advertiser, I’ve never found any benefit from Google extending my reach beyond what I have explicitly intended to target. Though, having the option to explicitly layer the options on top of each other makes perfect sense, which is an area in which Google has done well.

On the other side of the coin, I’ve asked (informally) several AdSense publishers what their thoughts have been about semantic extensions, remarketing, and audience buying. For the most part, they are in the dark about the changes Google has made of the last year or two, but they have noticed consistent decreases in their CPM (with the exception of game websites and other previously extremely low CPM sites), and ‘less relevant’ ad serving.  Hence, putting two and two together, a lot of my clients have connected directly with publisher and cut out Google (and its 32% fee) altogether.  Back in the early days, these direct buys were the main way to buy inventory. Google changed that, and the DSPs added to the fire. I wouldn’t go so far as to saying the tides are rolling back to those original 1-to-1 relationships, but I’m seeing more of these relationships occur, and I wonder when/if Google will take notice.

The answer to me is simple; keep semantic targeting based on semantics, and only serve ads in the manner advertisers explicitly intend to serve them.

 
Opinions expressed in the article are those of the guest author and not necessarily Certified Knowledge. If you would like to write for Certified Knowledge, please let us know.
 

Google recently announced that they were going to change how ads rotate. In the past, if you choose ‘rotate evenly’ then the ads would rotate throughout the ads in an ad group. In the new setting, ads will rotate for 30 days and then Google will automatically change the setting for that ad group to Optimize for Clicks.

The New Ad Rotate Problem

There are a few big issues with this change:

• If you don’t get enough data in 30 days; you can never test your ads.
• This is a common problem is you are testing small accounts or use cross ad group testing techniques
• The best CTR ad might not be the best profit ad
• If you get busy and forget to finalize a test, Google will make the decision for you (I believe that ‘optimizing accounts means taking control from Google so you can make all the decisions).

Now, rotate never was 100% accurate before due to how Quality Score and ad serving worked; however, it was directionally accurate and allowed for testing.

When Rotate Is Reset

After talking with a lot of people and some insiders at Google, I finally have an answer to when rotate is set:

• When a new ad is added to an ad group
• When an ad is unpaused

The biggest issue is that you cannot see what ad groups are on ‘optimize’ and which ones are on ‘rotate’. There’s no visibility into the ad serving by ad group. The best you can do is look at ad serving percentages within the last 30 days.

Fixing the Ad Rotate Problem

If you’re an API user; you can just fire off a command to pause ads for a few minutes each week and then unpause them and all will be reset. However, most people don’t access AdWords through the API, so I wanted a solution that would work for non-API users.

After some creative thinking, I came up with a solution:

Step 1: Use the AdWords labels to designate an ad in each ad group as ‘pause’

Step 2: Create an automated rule that pauses ads with the label ‘Pause’ an hour each week or month

Step 3: Create an automated rule that unpauses all ads labeled ‘Pause’ one hour after previous rule ran

The advantage of this method is that you don’t have to create new ads in an ad group.

The disadvantage is that you have to ensure you always have to be careful that if you choose an ad labeled ‘Pause’ as an ad testing loser that you delete the ad. If you were to just pause it, the rule will then unpause that losing ad.

Modification to the Fix

I posted an initial question about this method on Google+ and Chris Kostecki came up with a great modification to this plan. Instead of tracking all the ads in the current ad groups, do this instead:

• Create an ‘evergreen’ ad in each ad group with the label of Pause
• Unpause those ads for just an hour each week/month
• After an hour, pause the ads with the pause label again

The advantage of this fix is that you don’t have to keep track of ads.

The disadvantage is that you’ll have an ‘average’ ad running for an hour each week/month.

I’m a fan of this modification; and will most likely use it instead of my original proposal

Update on Rotation – 90 Days or Opt Out Forever

Google has changed their mind on this setting after a lot of vocal complaints. Rotate will now work for 90 days. If you would like to opt out of this change, you can fill out this form to try and opt out forever. If enough people opt out; they will add Rotate back as a feature (so go fill out the form).

If you don’t opt out; or are reading this post way to late to opt out; then feel free to just have the rule run every month. That way you will continue to have rotate forever even if you forget to opt out.

Get Back to Testing

It’s a shame that there has to be a workaround for this ‘feature improvement’. It makes me feel like we’re in the early days of PPC when you had to duplicate campaigns at different bid levels to manage the content network.

However, if you are serious about testing for Profits or doing cross ad group testing then using ‘Rotate’ is the best setting to use.

With this ‘fix’ it only takes a few minutes to deploy these rules and labels so that you can back to your normal ad copy testing routine.