AdWords Bid Management And Account Structure Mistakes To Avoid

On my one-year anniversary of working as an AdWords practitioner instead of a Googler, Brad Geddes interviewed me about Quality Score for his WebmasterRadio.fm show, Marketing Nirvana. He asked for my thoughts on AdWords now that I’m a user rather than someone building it, and I thought that was a great question and one worth covering in this month’s column.

I’ve always had plenty to say about the most common mistakes advertisers make, but now that I’m auditing accounts more in-depth and managing a few myself, there are some new mistakes I’ve seen that I think are worth sharing so that you can avoid them in your own accounts.

Bid Management Pitfalls

Handing the reins of bid management to Google or another third party has risks — but perhaps not for the reason I so often hear, which is that Google could use the information to their own advantage. I believe Google wants to use the information to give advertisers the results they want, and I am a big advocate of automated bid management because doing lots of repetitive math and keeping track of lots of variables is something computers are simply better at than your average account manager.

Automating bids is especially useful when you’re working with large accounts; any time you have more than a couple hundred keywords, managing all these manually is just too slow and tedious. But relinquishing all control to the algorithms can have unintended consequences, and it’s useful to understand the risks so you can decide what’s right for your account.

Monitor Your Top Keywords’ Bids Closely

An advertiser I was working with noticed that they lost pretty much all conversions on the one keyword that perfectly describes their business. This happened several months ago, and because they were spending so much time managing long-tail keywords, the drop-off on their money keyword initially went unnoticed.

Be careful about what automated bid management does to the bids of your top keywords. Here a keyword was bid down from the first page.

Only much later (during my audit) did they learn that this keyword had seen a week of poor conversion performance following a landing page redesign. The algorithm noticed it, too, and dropped the bid 60% — effectively taking them off the first page of results. They noticed the conversion rate decrease and quickly fixed the landing pages, but they never looked at resetting the bids for their keywords, so it continued to linger on page 2 of the results. Had the bid algorithm given the keyword another chance, it would have been restored to the first page of results and they wouldn’t have lost a ton of sales.

The takeaway here is that it makes sense to manage head keywords differently than long tail keywords. Even if you put them on automated bidding, you must monitor these keywords closely so that any change in performance will trigger a warning and you can investigate.

AdWords could make this much easier if they allowed advertisers to add charts of keyword-level detail to the interface. Until that time, we have an AdWords Script that builds a nifty little AdWords dashboard in Google Sheets.

Build a dashboard for any campaigns, ad groups or keywords in your AdWords account in minutes using an AdWords Script.

Changing Bids Can Change Which Keyword Is Served

The second issue with bid management is that your carefully calculated bids may not be the ones Google actually uses. Because of the vagaries of how Google selects the keyword from your account that matches the query, it’s possible that lowering the bid on a poorly performing keyword simply shifts its clicks to another keyword that still has a higher bid. This is exactly why we built the Traffic Sculptor – it helps find when Google gives an impression to a different keyword than what you intended. This way, you can be sure Google is serving the ad for your selected keyword, using its bid, ad text and carefully selected landing page.

Budget Constrained Campaigns May Not Work With Automated Bids

A third pitfall I’ve noticed with managed bids is that they may not always play well with budget constrained campaigns. On several occasions, I’ve seen that a campaign that uses bid automation depletes its budget too early in the day, even when ad serving is set to standard (meaning that ads are supposed to be shown throughout the day). Obviously, when the budget runs out by 10:00 a.m. and no conversions happen, it creates a potential vicious cycle — the bid management tool will eventually reduce bids too much and the campaign may entirely miss serving ads at the times of day when they would have performed best.

Account Structure Issues

Another common problem I’ve noticed is with account structure and keyword match types. I still come across accounts that combine display and search in a single campaign and accounts that have hundreds of keywords in an ad group — both not best practices in AdWords.

But one account structure issue that’s a bit more subtle (and that I hadn’t given much thought until this year) is that if you’re deploying a lot of broad match keywords in an account, it’s challenging to see trends or answer questions about why results change. That can make it hard to know if your strategy is panning out. As much as I like broad match keywords for their ability to help discover what users are really typing into the search box, they can make life for advertisers pretty hard….

Broad Match Keywords Are Liars

We’ve all heard that CPCs change based on seasonality or because Google introduces things like Enhanced Campaigns. But knowing how this impacts your account is trickier than just taking a look at the average CPC. The reason is that most advertisers use some degree of broad match keywords in their accounts, and stats for broad match keywords are averages that include data from each of the hundreds of queries that matched the broad keyword. We’ve all heard the adage that averages lie, and broad match stats are averages that obscure what’s really happening.

For example, when you change the bid for a broad match keyword, you’re changing the query mix because it becomes eligible for a different set of queries. The different queries have different CTRs, conversion rates, etc., so it quickly becomes a nightmare to figure out why a keyword’s performance has changed. Was it a competitor, a change Google made, an optimization you did, or simply the fact that your query mix changed? If your results are dropping and you can’t pinpoint the reason, it’s hard to correct the course. So unless you’re analyzing exact match keywords, you’ll have to dig very deep into the search terms report to try and come up with an answer.

To further complicate things, even exact match keywords may see sudden changes in performance because of the aforementioned issue where Google may shift the clicks to another, similar keyword with a better rank. In this case, the best way to understand what’s happening is to run a query report rather than a keyword report.

The Top 10 Reasons Why Your Website Is Losing Sales

The online retail game is not an easy one to win, but new research and analysis might help cast a light on the glaring problems causing your site to lose sales and customers.

Collecting user feedback and complaints from 400 websites over three years, Qubit has narrowed its findings to 10 main problem areas. The results paint a picture of a consumer with much higher expectations for retail websites and their user-friendliness.

Based on Qubit's research, here are the top 10 complaints that online shoppers have about their retail sites of choice.

1. Price

Price has been the top online shopper complaint between 2011 and 2013, and it's maintaining its spot. Qubit's research shows that consumers' biggest gripes about on-site pricing fall into either the "too high for a given product" or the "too high for me" category.

According to Nomis Solutions, consumers are 7.4 times more sensitive to price online than in-store, so it's imperative that your site's prices are competitive with the rest of your market and can offer something others cannot.

2. Product Range

Customers are taking their shopping online for many reasons — one of the biggest being the expansion product availability. Online shoppers want an easier experience discovering items than what they have in-store, and your site should reflect that desire.
Qubit suggests improving your site's product range by investing in site resources such as recommendation engines, dedicated lists for new and seasonal products, and "related item" groupings.

3. Size

Image: J. C. Penney

Many retail sites still don't have reliable or helpful sizing charts for their customers. Online returns are frustrating, and creating even a basic measurement guide can drastically reduce the amount of complaints you'll get from peeved shoppers.

4. Site Functionality

According to Qubit, if Amazon's checkout button was down for one hour, it could lose $2 million in potential sales. Basically, it boils down to this: If you have a feature on your site, make sure it works.
Best practices for maintaining customer relationships in the event of a site issue include being transparent when there are problems and providing an entertaining error message, such as Uniqlo's pictured below.

Image: Uniqlo

5. Latency

Considering how impatient and trigger-happy the average online shopper is, it's clear that site speed is a growing area for complaints. This quick-to-click mentality will only increase in the future, so it's imperative that your site fix any and all latency issues now.

6. Site Search

Image: JackThreads

Nothing pisses off a customer quite like a search bar that doesn't work at Google-like proficiency. To quell the rage, Qubit suggests improving and expanding your site's tagging system and adding advanced search features.

7. Stock Availability

Another recurring user complaint involves stock availability. Though this complaint has decreased since Qubit's last study, it could still pose a problem for your site, as customers expect a greater selection and variety when it comes to online retail stock.
According to Quibit, consumers frequently complain when items that are no longer in stock or discontinued are still displayed on the site.

8. Navigation

If your site isn't clearly labeled and easy to navigate, customers will leave. Make sure your popular items and sales are highlighted and your various pages are easy to find and navigate.

9. Discounts/Sales

One recurring complaint that Qubit has discovered is difficult-to-find discount code boxes. If you're going to offer sales and coupon codes, don't make them impossible to use, as it'll lead to shoppers leaving.
Many retail sites have begun implementing time limits on product purchases, which make the consumer less price-sensitive.

10. Imagery

Image: Macy's

If you're site isn't pretty, you're doing yourself a disservice. Eye-tracking studies show that web users scan first, then read. This means that product images are more important than ever.

According to Qubit, consumers prefer to see multiple images before making a purchase, and want to see clothing on models rather than simply laid out. Adding video content and multiple images can go a long way in getting a customer to make a purchase.

Use Google Analytics To Create Campaigns, Not Just Track Them

We all know that analytics are good for understanding how a visitor behaved on your website — what pages they looked at, what buttons they clicked or what keywords they used to find you. You can also use analytics to track campaigns like email blasts and paid advertising.  That’s pretty basic stuff.

What many beginning marketers and small business owners don’t know is that you can use analytics data to create an online marketing campaign.

Analytics can answer some of the key questions you might ask when you want to create a campaign: Who are you targeting? Where do they live? When should you start your campaign?

You can even pinpoint what content will resonate with your audience and combine that with the who, when and where. Using data to craft campaigns means you’re not just guessing — and guessing is gambling. I’m not saying there’s no place in marketing for the gamble; but sometimes, we need the safe campaign success to build us up to the point where we can take that gamble.

Collecting The Data

As we figure out what data to collect, I like to build a document that will start to take shape as an outline for an ad campaign. I set a time period that seems representative of the season I want to target. If I’m targeting Spring Break Travelers to Florida, I look at data from that same time period over the last two to three years.
My major sections are:

• Timing Target: What time of year are you looking to book, sell, etc.? (If there is no timing target, just leave this alone.)
• Top Converting Traffic Sources: Look at the All Traffic report under the Acquisition tab in Google Analytics.
• Best Converting Landing Page: Look at the Landing Pages report under Behavior –> Site Content and view it with Ecommerce info sorted by transactions descending.
• Highest Visit Count Landing Page: Look at same report as above, but sorted by Visits descending.
• Top 3 Keywords Driving Conversions: This is trickier because we live in “Not Provided” land. You will need to use PPC or Webmaster Tools Search Query data to help you get an idea of keyword phrases that convert.
• Top 3 Converting States/Cities/Metros: Under Audience –> Geo –> Locations, look at cities or states and Ecommerce, then sort by transactions descending.
• Male/Female Conversion Ratio: Under Audience, look at Demographics and then Gender. Apply the Ecommerce info and sort by transactions descending to see who converts more.
• Top Converting Age Group Demographic: Under Audience, look at Demographics and then Age. Again, apply the Ecommerce info and sort by transactions descending.
• Purchase/Booking Window: This is somewhat difficult to gauge in Google Analytics. I recommend looking at Time to Purchase under Conversions –> Ecommerce. The tricky piece is attribution — if you are looking at “Last Click Attribution,” which is the default and where many beginners or small business owners start — then a majority of your conversions will be on Day 0. The reality is that, for a novice or beginner with analytics, this number is likely something you know from dealing with customers and sales or from an internal reservation or sales software you use.

I then take this data and build a campaign. I start with source. A campaign that focuses on Search Engines as a source begins with a different strategy than a campaign that focuses on social media or email marketing.
Once I’ve determined source, I look for content. I look at landing pages, keywords, and demographics to flesh out the best targeted content for my campaign. My last step is to determine how far ahead of time I need to promote this campaign (if at all). If you’re targeting a holiday or a specific season and your purchase or booking window is long (say 30+ days), then you need to plan your campaign timing accordingly.

Let’s build a campaign with example data now:

Metric	Target
Timing Target	Spring Break – Feb 20 to March 31
Top Converting Traffic Source(s)	Email Marketing
Best Converting Landing Page	/rental/blue-swayed-shoes/
Highest Visit Volume Landing Page	/rental/
Top 3 Keywords Driving Conversions	key west vacation rentals, vacation rentals in key west, key west rentals
Top 2 Converting States/Cities/Metros	New York, Ohio, Pennsylvania
Male/Female Conversion Ratio	7:10
Top Converting Age Group	35-44
Purchase/Booking Window	35-45 Days

Now, I’m ready to build my campaign and most of my questions are answered — all using analytics. I can build an email blast with information that caters to women who like a 3-bedroom rental in Key West. Likely, they have a family and are looking to ship about 45 days or so from arrival. Knowing my purchase windows tells me when to start my campaign. This one in particular should be online by January 20 at the latest and run until early to mid-March.

If the best converting rental is already booked, my job is to find rentals nearby that have a similar size, look, feel, location and cost. This allows me to book similar units with a fairly targeted strategy. If you use a CMS software or collect customer demographics, you can segment your email list by age, gender and location. We use PARDOT, but there are others out there that will do a great job, as well.

Now, you have a model and a direction in which to go to start building your first online marketing campaign that originates outside of the “create an email blast for everyone” or “put up some PPC ads.” Your campaign could target Social Media, or even a specific platform such as Pinterest. Successful campaigns are built on what you know about your audience and how you can craft your “story” to hit home with them. Using Analytics to collect this data should get you started in the right direction.