SEO data and APIs: what you need and what it costs

Full transcript

1,062 words · the narration as spoken in the episode.

Intro

Most SEO tools cost somewhere between one and five hundred dollars a month. And most of what they show you, you'll never use.

At the same time, the data that actually answers your questions is either free, or costs a fraction of a cent per request. The trick is knowing which data you need for which question.

In this episode I'll walk through the minimal data stack for the three-part method from episode one, what each part costs, and the habits that keep the bill small.

Start from the question, not the tool

Quick reminder of the framework. Part one, traffic: what is happening on my site? Part two, SERP: what does Google actually show? Part three, gap: what do competitors have that I don't?

Each question needs different data. And that's the whole point of this episode. If you buy data before you know the question, you pay for dashboards. If you start from the question, you usually need surprisingly little.

So let's fill in the boxes, one part at a time. For each one I'll say what the source is, what it costs, and what it can't tell you. That last bit matters as much as the first two.

Part 1 data: your own site, for free

For traffic analysis, the core data is free, and it's yours.

Google Search Console gives you clicks, impressions, click-through rate and average position, per query and per page, per day. It keeps about sixteen months of history. Through the API you can pull it every day into your own database, which means you stop losing history after those sixteen months. We'll set that up in episode four.

What it can't tell you: anything that happens after the click, anything about searches where you don't appear, and every individual query. Some queries are hidden for privacy. Episode nine is about exactly that gap.

Google Analytics 4 covers what happens after the click. Sessions, engagement, conversions, and which hostname the traffic arrived on. Also free, also available through an API.

What it can't tell you: which search query brought the visitor. Analytics doesn't see queries. That's why we join it with Search Console on the page level, in episode five.

One warning that I learned the hard way. Analytics counts everything that loads your tag, including bots and traffic on old domains that still point at the same property. Treat it as raw material, not as the truth. We'll filter it properly later in the series.

Cost for part one: zero. Your time, and a bit of setup.

Part 2 data: the live results page

For SERP analysis you need to see what Google shows, for the searches you care about, in the location you care about.

You can do this by hand in a browser for a handful of keywords. But your own browser is personalized and localized, and you can't store or compare it. So for anything repeatable, I use DataForSEO, which I'll shorten to DFS from here on. It returns the full results page as structured data: every ranking URL, its position, its title, and every feature on the page.

What does it cost? In my own billing, a live results request with the top ten results costs about three tenths of a cent. Asking for the top thirty costs about twice that.

So a check of a hundred keywords costs around thirty cents. Which means the real cost isn't money. It's discipline: deciding which keywords deserve a check, and how deep you need to look. For most questions, the top ten is enough.

Part 3 data: demand and competitors

Gap analysis needs two more things: how many people search, and what competitors rank for.

Search volume. DFS can return Google's own keyword planner volumes. One task can hold up to a thousand keywords, and costs about nine cents. That's the cheapest data in the whole stack, per keyword.

Competitor rankings. This comes from DFS Labs, which is their own database of rankings, built from their crawls. Labs charges a small amount per task, a bit over one cent, plus a tiny amount per row returned. A thousand rows lands somewhere around thirteen cents.

What it can't tell you: today's truth. Labs data can be several weeks old. When a decision depends on a specific position, check it with a live results request from part two. That's a third of a cent well spent.

A realistic monthly budget

Let's put that together for a made-up coffee equipment shop, example-shop dot test. These are illustrative plans, but the unit prices are the ones I've actually been billed.

A weekly check of two hundred keywords at depth ten: about two dollars and forty cents a month. A monthly volume refresh for a thousand keywords: nine cents. And three competitor exports of a thousand rows each, once a quarter: around forty cents, when it happens.

That's roughly three dollars a month for a site with a serious keyword set. Not three hundred.

And notice what's not on the list. No daily rank tracker for thousands of keywords. No backlink database subscription. No all-in-one dashboard. You can add those later if a specific question needs them. Most of the time it won't, because the questions in this series are answered by your own data, the live results page, and a volume number.

Cost-control habits

Cheap per request doesn't mean cheap by accident. Four habits keep it that way.

Validate first. Before a big run, send one small task with a few keywords and check the result. For example, both the singular and the plural form of your main term. If the form is wrong, the whole dataset is wrong. That test costs nine cents, a big run costs much more.

Cache everything. Save the raw response from every paid call to disk before you parse it. If your script breaks, you parse again from the file, you don't pay again.

Choose depth on purpose. Top ten unless you have a reason. Deeper results cost more and are rarely needed.

And gate it. Every paid run should show an estimated cost and wait for a yes. When I let Claude Code run analysis scripts, that's a hard rule: no paid call without an explicit go. Episode six shows how.

Wrap-up

So: free data for your own site, a few dollars a month for the rest, and a gate in front of every paid call. The budget calculator for this episode is linked in the description. Next episode: DataForSEO from zero, from a new account to your first API call. See you there.