When is SEO a Meaningful Growth Strategy?
And the changes we made at Railway to make an immediate 30% impact on search - both traditional and AI chat.
I've been thinking a lot about SEO lately - it obviously affects AEO (Answer Engine Optimization), but the two are often conflated. More so, when does SEO actually make a business-changing impact?
Here’s some rhetoric I’m seeing (exaggerated, sure, but paints the picture of directionally incorrect advice I’ve seen):
SEO is always important
SEO = AEO; they are one in the same
Many efforts being lumped into the AEO umbrella
Let's be Clear: AEO ≠ SEO
Before we dive into what actually works, let's untangle some confusing words. AEO (Answer Engine Optimization) is bigger than SEO. Way bigger.
Yes, SEO impacts AEO - if you rank well in search, you're more likely to show up in AI-generated answers. But that's just one piece. If you do the following, you’ll also show up in AI answers:
Have a Reddit engagement strategy and execute on it
Produce enterprise media reports
Appear in dev blogs
Be mentioned basically anywhere credible information lives online
SEO for the last 20 years has been very Google-centric. As soon as AI chat enters the conversation (no pun intended), we have more ways to interact with information on the internet. Our pond just became an ocean.
While SEO strategies help you rank for specific keywords and topics like “deploy PostgreSQL”, AEO determines whether you show up when someone asks ChatGPT "How would I deploy a PostgreSQL database?", with follow-up questions like “Is this the most cost-effective solution?”. The first is about keywords and page structure. The second is about being part of the broader conversation.
There are many ways to be part of the broader conversation - whether it’s with the end user vs buyer, in earned media, on your own pages, on socials - the list goes on. One could argue this is all “SEO” - but that’s like saying “if you’re present on the internet, you have good SEO”. That’s not right either.
So when does SEO, in the more traditional sense, have its own legs to stand on as a growth strategy?
Redefining SEO in an AI-native World
There are two pieces of SEO: technical SEO involves making sure the information you already have is easily discoverable and searchable. The rest of SEO or what some might say is content SEO, involves putting out new content to make your presence bigger.
Technical SEO involves making changes to things like your site structure, raw HTML, breadcrumbs, etc. Content SEO is what it sounds like - putting out new content.
Many traditional SEO audits focus on a long list of tiny optimizations - like tweaking meta descriptions. It is extremely difficult to do these at any sort of scale without thinking programmatically. Why would you spend 2 weeks getting a 2% lift on traffic? Not worth it. So what is?
The real opportunity in SEO isn't in these incremental improvements. It's in thinking bigger about what pages you create in the first place, how they are programmatically structured, and how you can make one change to affect hundreds or thousands of pages.
The Zapier Playbook Everybody Talks About
Zapier’s SEO growth is classic story. They have many long-tail integrations, creating large surface area, and a prime use case for programmatic reach in search. When you're looking to connect Slack to Google Sheets, what do you search for? Not "workflow automation tool." You search "connect Slack to Google Sheets".
Zapier owns those searches for thousands of app combinations.
Each integration page follows a predictable formula:
"Zapier lets you connect [App A] with [App B]..."
Clear value prop about automating work
Actual templates you can use
Just enough content to rank without being spammy
This is the business-augmenting growth strategy, not the 2% lift you’d get from minor improvements.
SEO makes meaningful business impact when the possible reach goes for beyond your category reach.
Let me repeat it for the people in the back: if your SEO strategy is just focused on one category, you’re going to have an impact, but not an exponential one. This is what I’d consider table stakes SEO - if you’re trying to be the leader for a category, you have to show up in search, of course. However, your natural ceiling is how many people know about the category itself. Without something more, this focus is not going to 10x your business (even though it is, to be clear, necessary to not lose).
SEO as a growth strategy (rephrase: as a play-to-win, not play-to-not-lose) makes sense in cases where you can latch onto other problems or brands people already talk about, like through integrations.
My Practical Suggestions for SEO
So now I’ve hopefully convinced you that integrations could be your growth engine. Now what? Here's exactly how I bumped Railway’s Google Search traffic to integration pages by 30% within a week.
1️⃣ Audit technical SEO and programmatically fix
The important things to check:
headers
breadcrumbs
canonical links
meta tags
To be clear, all of these should be rendered programmatically in a search-friendly way. Once you start updating these manually, you become a hamster in a hamster wheel. Once you stop running (doing the manual work) - your impact fades.
2️⃣ Make the URL structure clear
Your URL structure matters more than you think. Going from /template/-NvLj4 to /deploy/fastapi makes the URL more legible, discoverable, and sets you up for even more programmatic pages to be created.
Yes, this means handling URL changes. Yes, it's a pain. Initially, you can use rewrites instead of redirects for a quick and dirty patch. Ideally, you store a trace of all name changes like Github does for PR names, or Linear for previous tickets created.
3️⃣ The content marathon
Here's where most teams fail. They try to update everything at once, burn out, and give up. Don't do that.
Instead:
Start with your top 20% of pages by visits
Use a simple structure that you can partially automate to get a first draft
If you have a community (we are extremely fortunate for ours), lean on them to make updates
There’s nothing groundbreaking here - but a start to making crawl jobs for Google (and OpenAI, etc) easier.
As soon as we shipped these changes, we saw a 30% bump in search traffic within a week.
While just the above won’t get you 10x returns, it sets up a foundation to build a programmatic strategy that will.
Measuring What Actually Matters
Many people look at rankings for individual keywords or topics. However, that only tells part of the story. You really care about clicks, especially in aggregate - so you don’t have to track hundreds of terms.
Specifically, looking at:
Total search traffic volume over time
Search traffic as a % of total traffic (ruling out growth from just general efforts)
Conversion from search traffic to signups
Tracking the negative is important, too, especially when rolling out new content pages. A false positive is rolling out poor conversion content pages - you get tons of traffic, but no conversions. This may in some cases be a signal the effort is misplaced on the wrong topics.
Thanks for reading - find me online talking about marketing, analytics, and broader tech - on LinkedIn, and on X.