Running SEO on the right operating system

Not every company can operate like Facebook.

Running SEO on the right operating system

Operating systems are the basic brain function of computers. Companies and their departments also run on operating systems: sets of beliefs driving decisions and processes.

Problems arise when the hardware and operating system aren’t compatible. You can run Windows on a Mac (whether you should is a different question), but Windows on an iPhone renders the hardware unusable. In the same way, many companies try to run a Growth OS like Facebook when they should actually run a Marketing OS like Salesforce. One of the core symptoms is driving decisions with data spite not having enough.

The solution is to identify the right OS for your hardware and make decisions accordingly. The best companies don’t slow themselves down by emulating a successful but incompatible company. They know what data they have and how to use it to make decisions fast and well.

Growth vs. Marketing

In his outstanding book “Everybody Lies”, Seth Stephens-Davidowitz tells the story of the Facebook news feed:

When Facebook introduced the news feed, users flipped out. However, data said they loved it, so it stayed: “Zuckerberg in fact knew that people liked the News Feed, no matter what they were saying in the groups. People were spending more time on Facebook, on average, than before News Feed launched. And they were doing more there - dramatically more. In August, users viewed 12 billion pages on the service. But by October, with News Feed under way, they viewed 22 billion.” (From “The Facebook Effect: The inside story of the company that is connecting the world.”) Engagement almost doubled through News Feed.

Facebook has so much data it understands users better than they understand themselves. They can make decisions purely on data because they have 3 billion users.

The majority of companies can’t operate on a Growth OS like Facebook because they don’t have enough users or customers to generate enough data. They’re doing Marketing, not Growth. The difference seems subtle, especially since so many people call themselves Growth Marketers these days, but it’s actually stark in contrast:

  • Growth = you know users better than they know themselves
  • Marketing = users know themselves better than you know them

As a result, Growth OS companies can focus on measuring as much data as possible to make better decisions. Marketing OS companies need to make decisions based on qualitative models fed by market research, post-sign-up surveys and interviews.

Quant models vs. qualitative data

I’ve yet to see a good attribution model for SEO for a company with a sales funnel. I argue it doesn’t exist. On the other hand, I’ve seen many companies where attribution and survey data don’t match at all.


If you have a sales-driven funnel, you usually go after larger but fewer deals. You’re more reliant on humans qualifying leads, entering the lead source correctly and maximizing pipeline opportunity (selling more).

Sales-driven companies deal with a lot of noise in SEO. Since the feedback loop is slow and top-of-the-funnel (TOFU) keywords have higher search volumes but lower intent, a lot of SEO traffic either doesn’t convert or takes longer than most attribution models measure (AI chatbots might change that). Customers will evaluate a $100,000/year product much more than a $1,000/year product, so your website gets a lot of seemingly non-converting traffic, but it just takes longer. Most attribution windows look at 90 days, maybe 120, if you’re lucky. How long is the average sales cycle for a $100K product? Often around 6 months.

Performance marketing, influencer marketing or email marketing, often convert faster (for various reasons), which is why the paid marketing team usually decides what attribution model the company uses.

Another problem is fewer but larger deals blur your view of what content or page type drives leads. I’ve seen a million Dollar deals coming to a glossary article from organic search. I doubt the article was the only reason for the company to become a customer. Still, technically it’s attributed to SEO and means the glossary drives millions in pipeline. So, should the company invest $100,000s into more glossary content to drive millions in pipeline? Probably not.

While most Marketing-led companies could run much better quantitative models, they should focus more on qualitative data.

The right OS for the right company

I distinguish between 3 business models setting the tone for either Product-led or Marketing-led SEO:

  1. Product-led (UGC, marketplaces, large e-commerce)
  2. Marketing-led w/ self-serve funnel (SaaS)
  3. Marketing-led w/ sales-driven funnel (Enterprise Saas)

The driving factor for a quant or qualitative model is TAM (total addressable market). The larger your customer base, the more data you can collect and the more you can let it drive your decisions. Meta’s TAM is everybody, and they can figure out the exact impact of the smallest change. But when you target a market of 1,000 companies, you need to operate on qualitative data.

A question to ask yourself when unsure about the right operating system is, “Does the data I look at tell a story without too much interpretation?” A lot of times, when I work with Marketing-led companies, and we look at which pages drive the most leads in SEO, the homepage and a few landing pages bring in 60-80%. It’s hard to make a data-driven decision with such a distribution. But when we ask customers how they first found the company, the majority say organic search. If, on the other hand, you run a large ecommerce site, it’s often easy to spot the SEO money-maker pages.

Making good decisions in Marketing-led SEO

My first recommendation for Marketing-led SEO companies with a sales funnel is expanding the attribution window to at least 180 days and evaluating channel performance on a multi-touch attribution model. Whether weighted, u-shaped or linear can be debated, but the idea that an expensive product is sold with a single touch is absurd. Customers have many interactions with the website, see the brand on many platforms and hear about it from others.

My second recommendation is to use interviews and surveys to drive more decisions. With tools like Grain (I’m an investor), Lookback, or User Interviews, you can run a regular program of interviewing existing and future customers. The key is having someone in your marketing team who takes ownership of bringing new insights to the rest of the org.

Lastly, a lot of Marketers would benefit from working like product managers and investing the time to exercise good judgment. Good judgment is the result of questioning your beliefs, seeking counsel from people who “have done it before”, and thinking through possible outcomes. It’s the art of framing a problem sharply and collecting information that helps you see how different decisions unfold. The key is writing all of this down, sharing it broadly, and archiving it for the future, so new teams don’t start from scratch.

Thanks to Nigel Stevens for being such a good thought partner on the topic.