Brand vs. Non-Brand — How to accurately assess your SEO performance


Web traffic from Search Engine Optimization (SEO) is a crucial component for most successful companies—how to split and accurately assess SEO performance

by Andrés Tapia and Ole Bossdorf

Web traffic from Search Engine Optimisation (SEO) is a key component for most successful companies. Yet many of them still struggle to fully understand what drives their SEO success. One reason might be a too simplified view on SEO — some companies just fail to acknowledge that SEO is comprised of brand and non-brand traffic.

At Project A, we always strive to understand the impact of our marketing efforts — thus we dove into this problem on how to split these SEO traffic types and how to accurately assess SEO performance.

Let’s imagine a hypothetical company, Shoes24, that sells shoes online. The marketing department of Shoes24 is divided by marketing channel (Facebook, Affiliate, SEO and Social Media), with a manager and a budget allocated for each of them. To have a better understanding of their marketing performance, they created a report which includes three main KPIs (clicks, revenue and costs) broken down by channel.

For the sake of simplicity, we didn’t include other channels like Direct, eMail, SEA and so on

The report is regularly used by the team to track marketing performance. Now imagine that the Affiliate and Facebook channels were merged in a single row called “Nonsense channel”.

How useful would this report still be for the Affiliate and Facebook team? As their two channels are mixed together, they wouldn’t be able to know if a sudden drop in revenue is caused by one channel or another.

Welcome to the daily life of almost every SEO Manager.

At most companies, the SEO report is just a compound of two very different traffic sources: brand traffic and non-brand traffic.

Even though all these examples end up as SEO traffic on your site, they are actually quite different. Brand SEO traffic is usually unrelated to your SEO work. Instead, this traffic type depends on brand awareness and investments in offline marketing (TV, sponsoring, physical banners etc.). Non-brand SEO traffic however actually depends heavily on your SEO investments such as content creation or website optimisation. The difference between these two types becomes more apparent when looking at it from a funnel perspective:

A simple solution to this problem would be to split the SEO report on brand vs. non-brand based on the keywords the users were searching, something that until 2011 was the standard approach (and easily done with Google Analytics).

The problem started when Google decided to discontinue connecting organic keyword data and transactional data in Google Analytics. Instead of seeing the keywords that brought traffic to your site you would see the almighty “(not provided)”, making it impossible to split the transactional data between brand and non-brand keywords.

The ‘not provided’ issue in Google Analytics

This leads to several issues:

  1. Performance evaluation: It’s difficult to assess the performance of SEO efforts targeted at increasing non-brand SEO traffic.
  2. SEO controlling: Algorithm changes, penalties and drops in non-brand rankings could remain undiscovered because of healthy brand traffic.
  3. Brand performance: The performance of offline or “brand channels” could be undervalued, as part of their traffic fall now under the SEO umbrella.
  4. Budget allocation: CMOs struggle to define SEO budgets as they are unable to assess their potential uplift.

Ok enough with the problems — let’s talk solutions

Next we will line out in 5 steps how we try to solve this together with our portfolio companies at Project A Ventures. In short: We identify brand keywords and measure their traffic share on each landing page. Then we apply these shares onto our attributed SEO conversions per landing page to calculate SEO brand & SEO non-brand conversions.

Step 1: Determining your brand keywords

Ideally you already have a list of keywords which are inevitably tied to your specific brand. These are terms where one can safely assume that the user is not looking for any non-brand search result but specifically for your company. Ideally you also add common misspellings to your brand keyword list which appear in your Google Search Console results.

Note: Defining these brand terms can become increasingly difficult for more non-brand brand names (e.g. Shoes Online Inc.) as the user intent is less clear.

Step 2: Calculate brand ratio per landing page (group)

The next step is to analyse how likely users enter your landing pages through non-brand versus brand terms. This information is readily available within search consoles of all major search networks (Google Search Console / Bing Webmaster Tools / Yandex Webmaster etc.). All you need is a breakdown per query for each landing page to calculate which percentage of the total clicks for this landing page goes to your previously defined brand keywords. This is how the breakdown looks like for Shoe24.

Calculating your brand / non-brand ratios

We calculate these ratios per week to ensure a sufficient amount of SEO traffic per landing page. In case of a large website it makes sense to group your landing pages after a certain level (e.g. /men-shoes/summer/slippers & /men-shoes/summer/flipflops both belong to the landing page group /men-shoes/summer).

Step 3: Figure out SEO clicks and conversions per landing page (group)

Next to this list of brand & non-brand ratios per landing page & week you also need to identify your actual SEO traffic and conversions per landing page based on your chosen web tracking tool. This information is not available on keyword level due to the “not provided” issue we discussed earlier. Still you can extract SEO clicks & conversions on landing page level e.g. from Google Analytics (Acquisition > All Traffic > Channels > Organic with landing page as a primary dimension).

However we recommend using a more sophisticated marketing data pipeline which extracts information from your web tracking tool and your backend. This pipeline enables you to run advanced attribution models and for example only considers paid orders or qualified leads.

After these first three steps you should end up with two datasets:

  • Brand / non-brand ratios per landing page (group) for a specified timeframe
  • SEO clicks, conversions per landing page (group) for the same timeframe

Step 4: Calculate brand & non-brand clicks per landing page (group)

Now you can join these datasets using the landing page as your primary key and apply your brand / non-brand ratios for each page and metric. What you potentially end up with is something like this:

While this is already quite close to your final result, we still need to consider the fact that usually non-brand searchers are further away from a buying decision than brand searchers.

Step 5: Account for different user intentions

While applying our brand / non-brand ratios to our SEO clicks is sufficient, you probably have already noticed that this doesn’t do the trick for our conversions columns. We discussed before that brand searchers’ intentions differ significantly from non-brand searchers. Someone who is actively looking for your brand is usually more likely to convert on your website. One way to find out if this is the case for your business is looking at conversion rates from your brand & non-brand SEA campaigns. If these conversion rates differ significantly you need to account for this difference by calculating a brand conversion rate factor:

brand conversion rate factor = brand conversion rate / non-brand conversion rate

We usually use brand & non-brand SEA campaigns as reference points to calculate this factor.

Let’s assume your brand SEA traffic is converting at 6% overall while your non-brand SEA traffic only does so at 3% overall. Hence the resulting brand conversion rate factor would be equal to 2. Here is how you use this factor when calculating brand / non-brand SEO conversions:

conversions = brand conversions + non-brand conversions

non-brand conversions = conversions * non-brand ratio /
(brand conversion rate factor * brand ratio + non-brand ratio)

Applying these formulas for every landing page (group) in a specific timeframe will definitely get you closer to really understanding your SEO performance. This is how our data pipeline looks like for these 5 steps:

Using the Solution

Putting this solution into place, either as a one-time analysis or as an automated report in your data warehouse, helped us in a variety of ways.

First of all, we were able to single out non-brand SEO traffic more effectively from brand SEO traffic in reports for our management. This helped us to realise that overall SEO traffic growth is mainly due to more non-brand SEO traffic:

More importantly, we also realized that most of our SEO conversions actually stem from brand traffic which is unrelated to our SEO investments such as content generation or technical website improvements.

Still we were able to see that the total share of non-brand SEO conversions is increasing over time, ensuring us that we are on the right path.

Furthermore this approach helped us to:

  1. Estimate the impact of building a new content section on our website: By backing our traffic estimations with revenue it was easier for management to accept our proposals and allocate the IT resources we requested.
  2. Isolate the impact of “brand channels” like sponsoring by adding up SEO brand + SEA brand + Direct in a single report: We then learned that our brand performance was dropping on a YoY basis.
  3. Evaluate the performance of our SEO team.

We hope this post helps you to get a better grip on your real SEO performance — maybe you are able to justify certain investments or emphasise to what extent your company relies on brand awareness. One thing we would like to tackle in the future is already distinguishing between brand and non-brand SEO clicks in our attribution modelling — we hope this delivers even stronger arguments to invest more in search engine optimisation.