Module 2 — BotPass 101: measurement, value, and decisions

📊In this module you make the leap that separates people who "have bot data" from people who "know

The pretty-data trap

There is a cinematic moment when someone installs BotPass for the first time. They open the panel, see charts, see bot names they recognize (GPTBot, ClaudeBot, PerplexityBot—), see pages listed with numbers beside them, and feel a small rush: "I can finally see what is happening." That rush is real and deserved, because until recently this was a total black box. But it hides a trap: pretty data is not decisions. Seeing that GPTBot visited your site 4,000 times last month is interesting, but it does not tell you what to do with your afternoon today.

The most common beginner mistake is staying in the contemplation phase. They look at the dashboard, are impressed, take a screenshot for the team, and then… nothing changes on the site. Months pass and the GEO Score stays the same, because looking is not the same as acting. The antidote is a rule you should tattoo on yourself: every metric you look at must have an associated action; if it does not, it is vanity. When you open BotPass, do not ask "what data is there?"; ask "what decision does this data let me make?" This entire module trains you to ask that second question.

The compass and the map: the two lists that change everything

There is an image worth fixing before you touch any button. BotPass is a compass: it shows you where bots and agents are looking, which content they are trying to use. But a compass alone does not take you anywhere; you also need a map of what matters to you as a business. GEO happens at the intersection of the two.

Imagine the two symmetric mistakes made when one of the two is missing. If you optimize pages that nobody — human or bot — consumes, you learn very slowly: you are polishing a corner nobody enters, and no matter how good it gets, no AI answer will change. But if you optimize only what has high bot consumption without asking whether that page is a source of truth for your business, you also learn slowly: you may be polishing an old landing that attracts many crawlers but does not represent what you want AI to say about you. Wisdom is in crossing both signals. That is why, throughout this module, you will always work with two lists in parallel: the consumption list (what bots read — data BotPass gives you) and the value list (what matters to your business — data you supply). Pages that appear on both lists are your treasure; that is where the work starts.

What to look at in BotPass without fooling yourself

When you open the dashboard, tame the urge to look at everything at once and organize around four questions — always the same, always in this order.

The first is which bots appear. Not all are equal, as we saw in Module 1: discovery and navigation bots come to read in order to answer, and therefore are the direct path to citations, while training bots have a slower effect. When you look at your bot list, the useful question is not "how many are there?" but "which of these can give me a citation soon?"

The second is which URLs they visit. Here you start to see the real geography of AI attention on your site, which almost never matches what you imagined. It is normal to discover that bots ignore your star page and instead devour a technical guide you had half abandoned.

The third is how often and how recently. A page bots visit daily and an hour ago is very different from one they visited once three weeks ago. Frequency and recency tell you which parts of your site are "alive" on AI's radar.

And the fourth is which part of the site they are taking: pricing, guides, documentation, policies. This grouping by content type is what connects consumption to risk, because it is not the same for them to take an opinion blog post as your pricing terms, where an error is costly.

"Estimated value": how to read it without lying to yourself

If your BotPass installation shows an "estimated value" per page, it is a useful signal — but only if you read it with a cool head. The golden rule is: use it to compare pages with each other, never as exact euros. Estimated value is not an invoice; it is a ranking dressed up as currency. It is useful to say "this page seems more strategic than that one," not to say "this page made me €37.40."

In day-to-day practice, what you really want to answer with that figure is which of three dimensions each page stands out in. Some pages are valuable for visibility, because they can bring you citations in your category. Some are valuable for precision, because they contain sensitive data — prices, policies, terms — where an error propagated by AI does real damage. And some are valuable for monetization, because they are collections that, if consumption is high, you could license or package (something you will see in Module 10). Classifying your pages by which of these three reasons makes them important is more useful than any absolute figure.

Popular is not the same as priority

This distinction is important enough to deserve its own section, because it is where most people go wrong. A popular page is one that receives many bot visits: typically the home page, category pages, navigation pages. They attract crawlers by structural position, not by content. A priority page is something else: one where, if a bot uses it to answer, an incorrect response is expensive. Pricing, policies, documentation, comparisons.

The temptation is to start with what is popular because the numbers are big and satisfying. But the strategic rule is the opposite: first make citable what is a source of truth. A pricing page with low bot traffic but data AI can propagate incorrectly is more urgent than a home page with huge traffic whose content is purely decorative. Popularity is volume; priority is consequence. Optimize for consequence.

The ninety-minute method that puts it all in order

Enough theory. Let us turn this into a procedure you can run today in ninety minutes and that leaves you with an actionable backlog. We break it into three chained steps.

The first step is to create your baseline. If you just installed BotPass, do not make decisions yet: wait at least a week for the system to accumulate representative data, because a single day can be skewed by any one-off event. If you already have history, use a fourteen-day window, which smooths odd spikes. With that window, take a snapshot of three things: your main bots, your main URLs, and, if available, value per URL. This snapshot is your "before photo," and you will thank yourself in Module 9 when you want to prove improvements, because without a before there is no before→after.

The second step is to build the two top lists: your ten pages by bot consumption and your ten pages by business value. The table below is the format to capture them; notice that the last column, "Source of truth?", is what crosses consumption with value and is what really matters:

The third step is to prioritize with a simple but disciplined rule. For each candidate URL, score four factors from zero to five: the GEO impact of improving it, the bot consumption it already receives, the risk if its content is wrong, and the effort to fix it. Then combine: Priority = (Impact + Consumption + Risk) − Effort. It is not a magic formula nor does it pretend to be; it is a way to force yourself to be explicit about why one page goes before another, and to avoid the bias of always working on "the most broken" instead of "the most worthwhile to fix." A broken page nobody consumes and that does not matter to the business can wait; a middling page that many consume and where an error hurts goes first.

The five actions that almost always come out of this analysis

When you run this method, the resulting backlog tends to fill with the same five families of actions, again and again, because they move the needle with the least effort. The first and most rewarding is usually adding a TL;DR with definitions and canonical links at the top of source pages: the ultimate quick win, low effort and high effect on extractability. The second is restructuring a page toward the answer-first principle, moving essentials up top. The third is consolidating duplicates when you discover the same truth lives on three different, slightly contradictory URLs: one truth, one URL. The fourth is creating a clean view for pricing or documentation when those pages are drowning in noise (a topic you will go deeper on in Module 6). And the fifth, when consumption of a collection is high and sustained, is starting to consider making it licensable (Module 10). You do not need to memorize the list; it will emerge on its own once you start looking at your data with the right questions.

Where BotPass is literally the solution

It is worth naming the typical problems BotPass solves directly, because they are exactly the frustrations people arrive with. When someone says "I do not even know if AI is reading my site," BotPass confirms bot traffic and which bots specifically, and that doubt disappears in minutes. When someone says "I do not know which pages to rewrite first," BotPass offers top URLs by consumption and, if available, by value, and the backlog orders itself. And when someone says "I am afraid of investing time in the wrong page," seven-to-fourteen-day baseline snapshots let you validate priority before spending hours — which is exactly what turns GEO from a bet into a process.

A complete example: from dashboard to an afternoon plan

To see the method in motion, imagine a realistic case. BotPass shows you that one bot visits /pricing and /docs/getting-started heavily; that another bot focuses on your comparisons and FAQs; and that among the most consumed pages appears an old landing with outdated pricing you thought was forgotten. That last line is a small alarm: AI is taking old prices and may be telling them to your potential customers.

The well-made decision, with the method's mindset, would look something like this. You declare /pricing the canonical pricing page and add a "last updated" line along with the terms, so it is clear which version is true and when it was set. You redirect or archive the old pricing landing to cut the error at the root. You add to /pricing an "FAQs by intent" section that answers the real questions people ask about cost. You put a TL;DR and clear numbered steps in /docs/getting-started so it is easy to extract. And, a detail many forget, you record the exact date of each change, because that is what will let you attribute in Module 9 whether the GEO Score rises.

That set of decisions becomes a concrete backlog, with estimated effort, that you can start today:

• /pricing — TL;DR + terms + FAQs — 2h
• /pricing-old — redirect to /pricing — 30 min
• /docs/getting-started — steps + requirements + example — 2h
• Changelog — record changes and date — 10 min

That — not staring at the dashboard — is "turning BotPass into decisions." Notice that none of those tasks comes from a whim: each responds to a panel data point crossed with a business priority.