What SEO Data Actually Costs, and What You're Billed For

Add a second person to your SEO tool and the price goes up, often by the whole plan again. Nothing underneath changed. Same crawls, same index, the same reports sitting in the same database — now readable by two people instead of one.

That is not a scandal. It is a pricing decision, and a normal one. But it does mean the number on your invoice is not a readout of what the data cost to produce, and if you are trying to work out whether a tool is priced fairly, that gap is the entire question.

So here is what actually costs money in an SEO tool, what tends to get billed instead, and — since it would be cheap to write this without doing it — our own numbers at the end, including the places where we are more limited than the tools we undercut.

The things that actually cost money

Fetching search results

This is the one genuinely per-unit, genuinely recurring cost in the business. A rank check is a request for one search result page, for one keyword, in one location, at one moment, and its entire value is freshness — a result from three weeks ago answers a different question than the one you asked.

The cost scales with keywords multiplied by frequency, and with nothing else. Two hundred keywords checked daily is around six thousand fetches a month. The same two hundred checked weekly is under nine hundred. Same tool, same team, same everything visible to you, and seven times the underlying spend — which is why choosing a tracking cadence is a budget decision disguised as a preference.

Maintaining a link index

The expensive one, and it behaves nothing like the first. To tell you who links to a domain, somebody has to have crawled a large share of the web and kept it current. That cost is almost entirely fixed: it is the same whether one customer asks or fifty thousand do.

Raw crawl data is not the hard part. Common Crawl publishes a fresh crawl of billions of pages and gives it away. What costs money is doing it continuously, resolving it into a link graph, and keeping it fresh enough that a link earned last week shows up this week. Every tool pays for that separately, which is the mechanical reason no two authority scores agree.

Crawling your own site

A site audit is per page, per run, and bounded by numbers you control. It is the easiest line to reason about and the one least likely to surprise anyone.

Asking a model

The newest line and by far the most expensive per request. A single rank check costs a fraction of a cent. A report that asks models what they say about your brand, across platforms, costs a few hundred times that. Nothing about it caches well either, because the answer is the product.

And the one that barely costs anything

Storage. History retention gets sold as a tier feature — thirty days here, two years there — and object storage is now priced in cents per gigabyte-month. A year of rank history for a thousand keywords is a rounding error against a single day of fetching them. When retention sits on a pricing table, it is there to separate tiers, not to recover a cost.

What you are billed for

Line the billing model up against the cost model and most of it does not touch:

Value-based pricing is not a crime and nearly every software company does it. But the bottom two rows do something worse than capture value: they make the bill unpredictable on purpose. A credit is a unit the vendor defines, can redefine, and never has to convert publicly into anything real. An overage tier means the vendor’s best month is the one where you guessed your usage wrong.

Neither is required by the cost structure. You can ration a per-request cost with a flat monthly cap that simply stops. Overages exist because they are more profitable than stopping, not because the data underneath arrives in credits.

The number that is never on the pricing page

The biggest single lever on what an SEO tool costs to run is its cache hit rate, and no vendor publishes it.

Two customers researching the same head term is one lookup, not two. Head terms overlap enormously across customers; the long tail overlaps almost not at all. So the marginal cost of any given query sits somewhere between zero and full price, decided entirely by whether somebody already asked it.

Two things follow. The first is that cost per customer falls as a tool gets more customers, on exactly the queries most people run — which is why “unlimited” can appear on a pricing page at all, and why it is always paired with a hard cap somewhere else, on the part that cannot be cached. The second is that a cheap tool is not automatically a reckless one. Most of the price difference between tools in this market is not data cost.

This is the part we built the billing around rather than the marketing. Reachavo sells a flat monthly research quota, and a repeat lookup served from cache does not count against it, because serving it cost us nothing. No credits, no overage line, no top-ups. When you hit the cap the tool stops making fresh calls until the next cycle instead of billing you for them.

How to price a tool before you buy it

Four questions that work on any vendor, including us:

• Price the thing you will do most, not the headline plan. For tracking, that is keywords times frequency. Write both numbers down before you compare anything, because that product is the only part of your usage with a real recurring cost behind it.
• Ask what a credit converts to. If nobody will state the rate in units you recognise — a report, a keyword, a crawled page — then it is not a price, it is a variable rate you have agreed to in advance.
• Separate the caps that cost from the caps that segment. Rows per report and fresh lookups per month track real costs. Seats, projects and retention mostly do not, which makes them the ones worth pushing back on.
• Take the free first-party data first. Your actual clicks and impressions come from Search Console’s performance report and from Bing Webmaster Tools, for nothing, and no paid tool can see them. A paid tool exists to tell you about other people’s sites, plus estimates on yours.

On those estimates, one thing worth internalising before you pay for precision: Google describes its own search volume figures as rounded averages over a period. Every tool in this market is presenting a modelled number downstream of data shaped like that. Coverage and convenience are what you are buying. Decimal places are not on offer from anyone.

Our numbers, including the bad ones

Reachavo is $19, $49 and $99 a month, or about twenty percent less annually. Flat. What that buys, and where it stops:

• 100 / 500 / 1,000 research queries a month. A hard cap that stops rather than bills. Cached repeats are free and do not count.
• 100 / 500 / 1,000 tracked keywords, refreshed weekly on the first two tiers and daily only on Scale. Daily refresh on a thousand keywords is the single largest recurring data cost we carry, which is why it is the top tier rather than a checkbox on every plan.
• 1,000 / 5,000 / 10,000 rows per report.This is a real limit and the one most likely to bite you. If your week involves pulling link profiles across dozens of domains, we are the wrong tool and the incumbents’ uncapped indexes are worth what they charge.
• 2 / 10 / 20 backlink reports a month, with competitor backlink gap on Scale only.
• 2 / 10 / 12 AI visibility reports a month. Scale gets fewer of these than backlink reports, which looks like a mistake and is not: each one costs materially more to run, so the quota is set from the cost rather than from the pattern of the other rows.
• 30 / 90 / 365 days of history and 1 / 1 / 5 seats. Neither is a cost line, by the argument two sections up. They are tier levers and we are using them as tier levers.

Behind all of that sits one number that is not on the pricing page: each plan has a fixed monthly data budget, and when a plan’s fresh calls reach it, the tool declines to make more until the cycle turns over. That is what “no overage” means mechanically. It is also the reason the row caps and report caps exist at all — they are how a flat price survives contact with a per-request cost.

The short version

Fetching search results and asking models questions are per-request costs. A link index is a fixed cost. Storage is nearly free. Seats, projects, retention tiers and credits are pricing choices sitting on top of that, and only rows per report and fresh lookups per month track the underlying spend closely enough to defend as pass-throughs.

None of which tells you a tool is overpriced — plenty of expensive software is worth it. It tells you which questions get you a straight answer. If our version of the answer sounds like the right trade, the plans and their limits are here, caps and all, with a seven-day trial on each.