The George Santos Guide to Search Volume (Ch. 1)

Why You Shouldn’t Trust Keyword Tools For Projecting Search Volume!

TL;DR: A new study warns against relying solely on search volume for traffic projections. Be careful when trusting popular keyword research tools as they can provide unreliable search volume estimates.

Watch the full webinar here:

Like most SEOs and agencies, whenever we work on projects, we use search volume as an indicator of how much potential traffic various keyboards could generate. 

We regularly cross-check this with Google Search Console data to verify the accuracy of search volume estimates so we can make better predictions and recommendations.

We’ve noticed some patterns that have made us very glad we do this for clients so we’ve done more expansive research to investigate it further. 

Our recent research demonstrates why it can be dangerous to blindly trust these tools for judging estimated search volume as the be-all and end-all of traffic predictions.

What is Search Volume?

You’ll likely already know that search volume is an estimation of how much a query is searched in a certain region per month. The purpose of using this estimation is typically to:

• Prioritize search queries
• Project impressions
• Project traffic (in combination with CTR)

By comparing estimated search volume to impressions, you can:

• Determine overall accuracy: How accurate is search volume for keywords you are ranking for?
• Find outliers: What types of keywords are not accurate?
• Find multipliers: What do you need to multiply your search volume by to get a semi-accurate projection

All of this should allow you to be able to have more confidence in your ability to forecast impressions and search volume both for yourself or for communicating initiatives with leadership.

Our Search Volume Research Analysis 

We pulled the data for 10 brands across different industries and compared their top 100 keywords data from GSC monthly impressions with the estimated monthly search volume predicted by Semrush and Ahrefs.

We then created a multiplier so we could measure and compare the effectiveness of both tools against GSC. In our analysis, we included brands from these categories:

•  Home Services
• Car Sales
• Home Improvement
• Automotive Directory
• Consumer & Commercial Printing
• Local Directories
• Medical rehabilitation
• Auto Service Multi-location
• Navigation Services
• Film & Television Review

You can see the overview of our findings and their multiplier below.

Overview of the Findings

• On average, Semrush more closely matched search impressions measured by GSC.
• Both tools were more inaccurate with “near me” queries although Ahrefs was slightly more accurate than Semrush on average.
• The industry category that was least accurate was an Auto service muli-location brand. The multipliers to equal SV for Semrush & Ahrefs were 30.95 and 62.31 respectively.
• The industry category that was most accurate was in general was car sales and automotive directories.

Our analysis provided a few key best practices that should help you better understand where these tools are unreliable and how to address it.

Regularly check your search volume against impressions

We have found that in some industries, search volume is almost a perfect indicator however in other industries or other types of keywords–it is way off. We’ll start with the more inaccurate categories and how to handle them then cover where we found they are most predictive of actual search volume.

“Near me” queries

While Ahrefs was typically better for “near me” queries, they both struggled and were consistently different from the search impressions measured by GSC. 

This is likely because “near me” search volume is the sum of all city-specific searches for one keyword, which makes it harder to estimate and collect reliable data.

When you are using “near me” search volume, consider what percentage of the US population are you targeting? 

For example, if you are targeting three states, then you’ll want to divide the population in those three states by the total population of the United States. 

This will give you a multiplier you use for “near me” keywords and it will give you a more accurate way to forecast when using search volume.

Sanity Check the Data for Extremes

Look for outliers especially with keywords where you are expecting lots of search volume and are getting very low search volume as well as for keywords where you're expecting very little search volume, but are getting a lot.

For example, our research showed keywords having a discrepancy as large as a factor of 396.48 when we compared projected search volume to impressions.

In general, we found that SEMrush seems to overestimate search volume while Ahrefs seems to underestimate search volumes on average.

Our approach is a preference to underpromise and over deliver so we tend to reduce Semrush projection by 20-30 percent.

It’s best to sort the search volume by largest to smallest and then check the top few and the bottom few whenever you're doing larger lists of search volume analysis. 

There are situations where words can mean multiple things and the intent of what is being shown is not the keyword intent that you are going for, so the search volume will be way off the mark.

Branded Terms

In almost all of the 10 brands we researched, both Semrush and Ahref overpredicted the branded terms–in one case it was 58 times the actual impressions.

When you are trying to rank for keywords that contain your own brand—or more importantly–other people's brands such as “[Car Brand] dealer” or “[Car Brand]  dealership”, these tools are likely going to guess way too high.

If you are using Semrush or Ahrefs for branded terms, then it’s best to check your keywords and create a divider to get a more accurate picture of what the search volume will be.

Queries That Contain The Cities

One finding that was surprising was the effectiveness with which the keyword research tools predicted queries that contain a city. Especially after seeing they were quite inaccurate for “near me” searches. 

Industries Where They Are Most Accurate

When you exclude “near me” queries the Automotive industry brands had volume estimates that matched their actual search volume closely.

For example, one of the car brands we researched had the majority of the tools within 10 to 20% on average. 

It seems for this industry, you can use Semrush and Ahrefs search volume with a lot higher degree of confidence.

The e-commerce industry seems to be another one where they're really good on average for both tools.

Best Practices For Better Search Volume Forecasting

To figure out how effective search volume is at projecting a client’s traffic we compare their top keywords, where our clients rank on the first page, and look at the last 12 months of average monthly impressions vs. search volume in Ahrefs and Semrush. 

When you are summing search volume, check if any keywords have the exact same volume estimate. If they are too similar to each other, they might be counted as the same and shouldn't be added to your total.

 If you are unsure, you can double-check that they have different SERPs. 

To repeat our process:

1.  Export Search Console data: All data you have and divide by amount of months.
2. Collect search volume: Pull SV out of your favorite SEO tool
3. Find multipliers: Calculate what you need to multiply search volume by to get to impressions.
4. Find multipliers for keyword categories: Calculate what you need to multiply search volume by to get impressions for keywords in specific categories.

Check The Limitations Of All Your Tools

While using this method can help you check and mitigate the limitations of Ahrefs and Semrush, no data tool is perfect, including Google Search Console. Patrick Stox, Technical SEO at Ahrefs explains:

“The GSC data can be used, but you can’t completely trust it. Rank trackers, missing data, click manipulation. There’s a lot of noise there. It’s actually a fun and complicated problem. 

No matter the data source, there are always some issues with it which is why you’re seeing more custom models. We take all the data and try to make sense of it.”

Google search console can have other issues as well. For example, Joy Hawkings of Sterling Sky found that Search Console seems to miss clicks from the 3-Pack on Mobile.

While tools continue to improve, it’s a good idea to verify the reliability of the data you are using for any tool as a best practice. Non-verified search volume should be used mostly for prioritization, not for estimates. If you are trying to rank for keywords where you have no impressions/clicks, collect data from multiple tools.

Key Takeaways from our Search Volume Research  

• Regularly check your search volume against impressions
• Sanity checks the data for extremes
• Branded terms can be way off!
• Queries that contain the cities were accurate while “near me” were not
• Check the limitations of all your tools
• Remove outliers from your projections