How to Tell if You're Wasting Money on Google Ads

How to Tell if You’re Wasting Money on Google Ads

Search advertising is the biggest expense for most digital advertisers, yet many are uncertain they are spending efficiently. Google AdWords (Now called Google Ads) and Bing Ad Center have evolved dramatically over the past several years, developing more features that enable advertisers to make their ad targeting more precise and their keyword bids more efficient.

This is great news in concept, but the practical complexities of this opportunity can be overwhelming. Most advertisers lack the tools needed to take advantage of the new Google Ads features. Say, for example, you have a modestly simple program with 5,000 keywords and 10 retargeting lists. That would yield 5.2 billion potential bid variations to calculate on a daily basis.

(5000 Keywords * 10 Audiences * 24 Hours * 7 Days * 210 Metro Areas * 3 Device Types = 5.2 Billion)

If you bid effectively, you can achieve extreme precision — eliminating wasted spend and producing huge returns. So how can you tell if you are best leveraging the tools available to get to the right bid for each scenario?

This step by step guide will help you quantify the amount of opportunity you have to improve your returns on Google Ads by pulling these levers.

But before you dive into the exercise, here’s a checklist of basic bid optimization questions you’ll need to answer in order to get started:

  • Do I understand the impact each keyword has on my deep funnel conversions? (Think leads, LTV, Lead Score, Phone Call Conversions, Revenue, Profit, Store Visits, Signed Contracts, Converted Pilots)
  • Can I bid on each keyword based on its ability to drive those results, or am I stuck with grouping keywords together and dealing with averages?
  • Can I bid on keywords that have little or no data? What about keywords with infrequent, high-value conversions?
  • Location: Am I making full use of location-based bidding? Am I bidding the right amount given which city the person is searching from? Do I account for daily trends within each city?
  • Day Parting: Am I bidding the right amount for the time-of-day/day-of-week combination?  For example, if my conversion rates are better on Thursday afternoon than on Wednesday afternoon, am I using that information to my advantage?
  • Device: Am I targeting specific devices or bidding differentially based on Device?  Do I know how much device matters?
  • Day of Week: Do I have a scalable way to bid differently based on Day of Week performance?
  • Audiences: Do I understand my audiences, and do I have a scalable way to bid the right amount for each audience based on performance? Am I updating my audience targets frequently based on Day of Week or seasonal trends?

If you can only answer yes to two or three of these questions, then you are a very typical advertiser. I talk to dozens of search advertisers every month and we almost always find that they aren’t fully leveraging the majority of these capabilities. And these are very smart advertisers, running sophisticated programs. They simply lack the technology to control this complex set of options.

So, are you ready to roll up your sleeves and find out how your program is doing?

For this, you will need:

  • Approximately 45 minutes of free time without distractions
  • Read-only access to your Google AdWords account(s)
  • Three months’ worth of Conversion or Revenue data in your Google AdWords account(s)
  • The ability to run pivot tables in Microsoft Excel (instructions are provided in this article).

In the rest of this article, I will take you through each step to evaluate your Google Ads spending. No matter how basic or complex your PPC program is, you’ll have all the tools you need to work through and discover your wasted ad spend. In our first analysis, we’ll take a look at keyword level efficiency.

Google Ads Analysis 1: Keyword Level Efficiency

Brief Description: For our first analysis, we’ll reveal the percentage of spend for non-converting keywords over a three-month period.

Industry Benchmarks

First, you need to make a judgment about where you should be. This one can be quite fuzzy, as it varies dramatically based on conversion type. Here are some trends and recommendations we make based on business type, industry and other factors:

  • B2B companies with high value, low volume conversions: it’s common to see 60% of spend applied to keywords without conversions.
  • B2C Retail companies with low value, high volume conversions: It’s common to see 15-30% of spend applied to keywords without conversions. For super high volume/ low-value conversions, we recommend using a ROAS threshold instead of conversions.  For example: If 2X ROAS is your profitability threshold, then run this analysis for percentage of spend applied to keywords that are below 2X ROAS instead of percentage of spend applied to keywords without conversions.
  • Automotive Companies where conversions are a Vehicle Description Page view, Request for Quote, Dealer Locator, Download a Brochure, Build and Price, etc.: It is common to see extremely low percentage of spend applied to keywords without conversions. In such cases, we recommend picking a specific conversion metric that you wish to improve (such as Request a Quote only). 
  • Travel Companies: We recommend looking at Bookings as a conversion.  We have some travel customers with hundreds of millions of keywords, and others with several thousand, depending on the business model.  If you have hundreds of millions, then we recommend applying filters and running this analysis on your top 100K keywords.
  • Financial companies where conversions are completed applications: We tend to see that the majority of spend is on head terms, and a very low percentage of spend is applied to keywords without conversions.  In this case, we recommend looking at percentage of the long tail keywords with clicks. Non-brand head terms are expensive in this space, so the more you can extract value from the long tail, the better.
  • Lead Generation companies that provide leads to local contractors: Performance numbers vary dramatically depending on the type of conversion. For initial form submission conversions, you should expect to see less than 10% of spend on keywords without conversions. You can improve the quality of leads by feeding deeper funnel conversions (such as accepted leads or booked contractors) into your bidding optimization program.
  • Publishers who make most of their revenue from Adsense or from running display ads: In this case conversions are typically extremely high volume/low value.  So we recommend using a ROAS threshold instead of Conversions. For example: if 2X ROAS is your profitability threshold, then run this analysis for percentage of spend applied to keywords that are above or below 2X ROAS.

Analysis Process

Once you’ve considered a benchmark for your program, you’re ready to start the analysis process:

Determine Accounts and Conversion Fields

  1. Determine Account: It is easiest to run this analysis on one account. If you have several accounts, then you could choose the largest account, or the account that best represents the rest of your program. Click into that account.
  2. Determine Conversion Field:  Navigate to your Conversions Page in Adwords (New Adwords Interface – Wrench -> Measurement -> Conversions). Determine which is the most meaningful for your analysis. I typically use “All Conversions” unless the account has a lot of high funnel conversions like product page views or watched a help video included.

Prepare and Download Data

    1. Navigate to the keywords tab within the account you have chosen to analyze
    2. Apply a filter: Campaign Type: Search Network Only. Note: If your report is longer than 1 million rows, you may want to add additional filters (such as Keyword Status: Eligible) because you are going to be downloading this to Excel.
    3. Change the Date Range in the top right corner to reflect the last 90 days
    4. Add Columns: Keyword, Impressions, Clicks, Cost, Match Type. Include whichever conversion column you decided to utilize. You may also want to include custom dimensions (such as brand/non-brand) if you want to see your results broken down as such.
    5. Download the report in Excel.

Excel Work

  1. Remove the summary rows at the top and bottom. Here’s how:

  2. Add a column at the end and name it “Clicks?” Then add another column at the end and name it “Conversions?” It should look like this:

  3. In the “Clicks” column, type the formula: =IF(ClicksCell, “Has Clicks”, “No Clicks”)

  4. Extend this formula to the entire column.
  5. In the “Conversions” Column, type the formula: =IF(ConversionsCell,”Has Conversions”, “No Conversions”)

[Note If you are using a ROAS threshold instead of Conversions, then replace “Conversions” with “ROAS Threshold.” Then create an ROAS column (Revenue/Cost). Use this formula to populate the ROAS Threshold column: =IF(ROASCell>[insert ROAS profitability Value],”ROAS GT Threshold”, “ROAS LT Threshold”)  In this example, the ROAS profitability threshold is 500%, so we used this formula: =IF(ROASCell>5,”Above 5X ROAS”, “Below 5X ROAS”)]

  1. Extend the formula to the entire column:

  2. Select all data and create a Pivot Table (Click “OK” on the next dialog)
  3. Rows: “Conversions?” (or optionally use “ROAS Threshold” instead)
  4. Columns: Match Type, brand, non-brand, etc (optional – depending on which dimensions you want results broken out by)
    1. Values: Cost (Sum of)
    2. Calculate the % of non-converted spend (spend with no conversions/total spend) for each column (broad, exact, total)

Analyzing Results

A healthy non-converted spend % is usually somewhere below 20%. Too low a percentage means too little keyword discovery. Too high a percentage means waste. Please see the industry benchmarks for your specific vertical. The account in the example screenshot has 6.3%. That is probably too low, meaning that efficiency is strong, but that there is an opportunity to explore the long tail for more volume. The next analysis will help us confirm that.

So how does QuanticMind solve this — and other — challenges?

Modifier Complexity: QuanticMind breaks out each keyword into distinct combinations of Keyword x Location x Hour of Day x Device x Audience (Audiences coming soon).  Then we determine the ideal CPC for each of these combinations. Finally, we use Machine Learning to back into each modifier to get the entire program as close to the ideal CPC as possible.

Hour of Day segments: QuanticMind solves the complexity by automating the creation of Hour of Day segments and automatically calculating bid adjustments for each segment each day, based on your business targets.

Long Tail Data Scarcity: QuanticMind uses Natural Language Processing (NLP) to bid intelligently on long tail keywords by automatically borrowing data from semantically similar head terms. This provides significant lift in terms of clicks and conversions from the long tail without the need for costly trial and error.

Location Data Scarcity for small cities: Similar to our practice with utilizing NLP for the long tail, QuanticMind compares city level demographic points to each other to allow small cities to borrow data from demographically similar large cities.  This provides lift by applying intelligent Location bid adjustments to cities that do not have sufficient data to support it.

Keyword Inefficiency: QuanticMind has the processing power to address each keyword separately, applying your business goals at the most granular level and generating incremental lift.

What’s Next

Thought this analysis was helpful?

This is just the beginning. We have four more analyses in store for you in our upcoming eBook, which will further help you determine if you’re wasting ad spend on Google Ads, as well how to create new efficiencies and uncover new opportunities to boost ROI and truly optimize your campaigns to their full potential.

Feel free to request a demo with QuanticMind. Bring us the results of your analysis and we can propose a solution to reach peak efficiency. Or ask us to conduct the analysis on your behalf.  We love this stuff!

Subscribe to the QuanticMind Blog
Please enter all required fields Click to hide
Correct invalid entries Click to hide

Seth Jennings is the Director of Solutions Engineering at QuanticMind. He has rich experience in the digital marketing space from his tenure at Oracle Marketing Cloud, BlueKai, and Adobe and is highly active with QuanticMind's clients.

Stay on top of performance marketing trends and innovations. Get our posts delivered straight to your inbox.