Demanda de rastreamento

O lado do "querer" do crawl budget — o que faz o Google querer rastrear suas páginas (popularidade, desatualização e inventário percebido), como a demanda encontra a capacidade do host e por que você não pode forçá-la.

Publicado pela primeira vez: 3 de jul. de 2026 · Última atualização: 3 de ago. de 2026 · Avançado

«Crawl demand is the 'want' side of crawl budget — how much a search engine wants to crawl a site or URL, as opposed to how fast it can (that's crawl rate/capacity). Google names popularity (links/PageRank), staleness (how often a page changes), and perceived inventory (how many URLs Google thinks exist, junk included — 'the factor you can positively control the most') as significant general demand factors — not a closed three-item formula; Google also points to site size, update frequency, page quality, and how a site compares to similar ones. Site moves temporarily spike demand. Host load acts as a ceiling on realized demand, not a driver — my synthesis: demand sets the priority order of URLs, capacity decides how far down that queue Googlebot gets. You can't set demand directly — you move its inputs by earning links, keeping content genuinely fresh, and cutting junk-URL inventory so quality signals feed back into the scheduler. Google is actually trying to crawl less overall while routing demand more precisely, so the real goal was never volume — it's correct prioritization. Most sites never need to manage this.» _(Tradução)_ (Síntese localizada do campo tldr em português para conferência na revisão nativa.)

«> TL;DR — Crawl demand is the want side of crawl budget; crawl rate/capacity

is the can side. Google names popularity (links / PageRank), staleness (how often a page changes), and perceived inventory (how many URLs Google thinks exist, junk included — “the factor you can positively control the most”) as significant general demand factors — not a closed formula; site size, update frequency, page quality, and comparative relevance also factor in. Site moves spike demand temporarily. My synthesis for tying it together: demand sets the priority order of URLs; host-load capacity decides how far down that queue Googlebot gets — Google doesn’t document this as a literal algorithm, but it’s the model that fits the evidence. A healthy server doesn’t manufacture demand, and high demand can still be capacity-throttled. You can’t set demand directly — the only levers are its inputs, and the scheduler turns demand up when quality signals from indexing improve. Meanwhile Google is actively trying to crawl less while routing demand more precisely, so the real goal was never volume — it’s correct prioritization. Most sites never need to manage any of this.

Evidência desta afirmação Google says Googlebot demand varies by site size, update frequency, page quality, and relevance compared with other sites; significant general demand factors are perceived inventory, popularity, and staleness. Escopo: large or rapidly changing websites Confiança: alta · Verificado: Optimize your crawl budget

» (Tradução) (Síntese localizada do trecho dezoito, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«## Crawl demand is the “want,” crawl rate is the “can” » (Tradução) (Síntese localizada do trecho dezenove, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«Google is explicit that crawl budget has two halves: the amount of time and resources Google devotes to crawling a site “is determined by two main elements: crawl capacity limit and crawl demand.” The way I frame it in my Ahrefs crawl-budget guide: crawl budget is “made up crawl demand which is how many pages a search engine wants to crawl on your site and crawl rate which is how fast they can crawl.” Demand is the want; rate is the can. Evidência desta afirmação Google describes crawl demand as one of the two main elements of crawl budget, alongside crawl capacity limit. Escopo: Google Search crawling. Confiança: alta · Verificado: Google: Large site crawl budget guide » (Tradução) (Síntese localizada do trecho vinte, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«This page is only about the want. The can side — the crawl capacity limit, the GSC rate slider that was removed in January 2024, how 5xx/429 responses slow Googlebot, and Bing’s manual Crawl Control grid — all lives on the crawl rate page. I’m not going to re-derive it here; when the two interact I’ll link across. » (Tradução) (Síntese localizada do trecho vinte e um, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«## The three demand inputs » (Tradução) (Síntese localizada do trecho vinte e dois, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«Google’s current guidance calls perceived inventory, popularity, and staleness significant general factors that drive how much it wants to crawl — it doesn’t present them as an exhaustive, closed formula. The same guidance also names site size, how often it’s updated, page quality, and how it compares to similar sites as factors Googlebot weighs. The three below are the ones Google explains in the most depth and the ones you can act on directly. Evidência desta afirmação Google's crawl-demand guidance discusses perceived inventory, popularity, and staleness. Escopo: These inputs influence crawling but do not provide a user-controlled demand setting. Confiança: alta · Verificado: Google: Large site crawl budget guide » (Tradução) (Síntese localizada do trecho vinte e três, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

A demanda decide quais URLs ficam na frente da fila; um servidor saudável apenas determina quanto dessa demanda pode ser realizado. Fonte: Google Search Central

A demanda de rastreamento ordena as URLs usando popularidade, mudança genuína e o valor percebido do inventário de URLs do site. A capacidade de rastreamento, baseada na velocidade de resposta, na estabilidade e nos erros do servidor, determina até onde o Googlebot pode avançar nessa fila ordenada. Uma infraestrutura mais rápida eleva o teto da capacidade, mas não cria demanda para URLs de baixa prioridade.

© Patrick Stox LLC · CC BY 4.0 ·

«### Popularity » (Tradução) (Síntese localizada do trecho vinte e cinco, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«“URLs that are more popular on the Internet tend to be crawled more often to keep them fresher in our systems.” More links and more PageRank pointing at a URL is a demand signal — it’s why your homepage gets crawled constantly and a deep, unlinked page barely at all. As I put it in my crawl-budget guide, “Popular pages, or those with more links and PageRank, will generally receive priority over other pages.” Internal links count here too: a page nothing links to (an orphan) has almost no demand working for it. » (Tradução) (Síntese localizada do trecho vinte e seis, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«### Staleness » (Tradução) (Síntese localizada do trecho vinte e sete, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«“Our systems want to recrawl documents frequently enough to pick up any changes.” Google learns each page’s rhythm. A page that changes constantly earns frequent recrawls; a page that never changes gets checked less and less. In my crawl-budget guide I describe the backoff Google applies to a static page: “if they crawl a page and see no changes after a day, they may wait three days before crawling again, ten days the next time, 30 days, 100 days, etc.” That per-URL recrawl cadence is really a crawl frequency question — I cover the mechanics there — but the underlying force is demand, and specifically staleness. » (Tradução) (Síntese localizada do trecho vinte e oito, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«### Perceived inventory (the one you control most) » (Tradução) (Síntese localizada do trecho vinte e nove, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«This is the headline lever. Google: “Without guidance from you, Google tries to crawl all or most of the URLs that it knows about on your site. If many of these URLs are duplicates, or you don’t want them crawled for some other reason (removed, unimportant, and so on), this wastes a lot of Google crawling time on your site. This is the factor that you can positively control the most.”

Evidência desta afirmação Google calls perceived inventory the crawl-demand factor site owners can positively control the most; duplicate, removed, and unimportant known URLs can waste crawling time. Escopo: large or rapidly changing websites Confiança: alta · Verificado: Optimize your crawl budget

» (Tradução) (Síntese localizada do trecho trinta, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«The subtle part is what it does to demand, not just capacity. It’s easy to think of junk URLs as “wasting crawl budget” — spending fetches on copies instead of new content. True. But there’s a demand-side effect too: a site whose knowable inventory is mostly low-value duplicates and parameter sprawl looks, to Google, like a lower-value site to crawl. Cutting perceived inventory doesn’t just free capacity; over time it concentrates demand on the URLs that deserve it. Faceted navigation, session IDs, infinite calendar spaces, and other spider traps are the classic inventory inflators — and the classic demand suppressors. » (Tradução) (Síntese localizada do trecho trinta e um, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«## Site moves and other demand spikes » (Tradução) (Síntese localizada do trecho trinta e dois, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«One demand driver isn’t about any single URL: “Additionally, site-wide events like site moves may trigger an increase in crawl demand in order to reprocess the content under the new URLs.” If you do a domain migration or a big replatform and notice Googlebot hitting you far harder than usual for a few weeks, that’s expected — Google has to re-fetch and reprocess everything under the new addresses. It’s a temporary spike, not a new baseline, and it’s a demand event competitor guides almost never mention despite it being verbatim in Google’s own documentation.

Evidência desta afirmação Google says site-wide events such as site moves may temporarily increase crawl demand so content can be reprocessed under new URLs. Escopo: large or rapidly changing websites Confiança: alta · Verificado: Optimize your crawl budget

» (Tradução) (Síntese localizada do trecho trinta e três, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«## How demand and host load interact: queue order vs. capacity gate » (Tradução) (Síntese localizada do trecho trinta e quatro, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«Here’s a mental model that makes the whole topic click — and I want to be upfront that it’s my synthesis, not something Google documents as a literal algorithm. It’s built on a Gary Illyes Q&A, quoted here through Search Engine Roundtable’s coverage: host load “sets a bucket of URLs in importance order and GoogleBot will crawl in that order based on the schedule the host load decided. If Google thinks your server can handle it, it will crawl the whole bucket, if not, it will stop.” Notably, per that same Q&A, host load tracks the importance of your pages — not the raw number of URLs you have or how many you want crawled. That page blocks automated fetching, so I haven’t been able to re-confirm the exact wording directly against the live source this pass — treat it as a well-corroborated paraphrase, not a verbatim primary-source quote. » (Tradução) (Síntese localizada do trecho trinta e cinco, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«Read that carefully and a relationship falls out — again, this is how I connect the pieces, not a mechanism Google has spelled out end to end: » (Tradução) (Síntese localizada do trecho trinta e seis, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«- Demand sets the order. The “bucket of URLs in importance order” is crawl demand — popularity and staleness deciding which URLs sit at the top of the queue.

  • Capacity sets how far Google gets. Host load / crawl rate decides how deep into that ordered bucket Googlebot actually crawls on a given day. “If your server can handle it, it crawls the whole bucket; if not, it stops.” » (Tradução) (Síntese localizada do trecho trinta e sete, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«So the two aren’t just multiplied together — they play different roles. A healthy, fast server doesn’t manufacture demand (it only raises the ceiling on how much of your existing demand gets realized), and high demand can still be capacity-throttled (a slow or error-prone server stops Googlebot part-way down the queue no matter how much it wants to crawl). This is why “I bought a faster server and Google still isn’t crawling my new pages” is such a common, frustrating result: capacity was never the constraint — demand was. » (Tradução) (Síntese localizada do trecho trinta e oito, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«## You can’t set demand directly — but the scheduler listens » (Tradução) (Síntese localizada do trecho trinta e nove, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«Short answer: you don’t set demand directly. You earn it, indirectly, through real links and real quality improvements that show up in indexing signals — nothing else moves it. » (Tradução) (Síntese localizada do trecho quarenta, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«There is no “crawl more” request for demand any more than there is for rate. But demand is dynamic, and Google has been unusually candid about how it moves. Gary Illyes: “If you want to increase how much we crawl, then you somehow have to convince search that your stuff is worth fetching, which is basically what the scheduler is listening to.” And the feedback loop is real-time-ish: “Scheduling is very dynamic. As soon as we get the signals back from search indexing that the quality of the content has increased across this many URLs, we would just start turning up demand.” The flip side too: “If search demand goes down, then that also correlates to the crawl limit going down.” (I re-checked these three lines against Search Engine Journal’s coverage this pass and they match verbatim; I still haven’t tracked down Google’s own podcast audio/transcript to confirm them as a primary source, so treat them as well-corroborated secondary quotes.) » (Tradução) (Síntese localizada do trecho quarenta e um, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«That reframes “how do I increase crawl demand” away from tricks. Fake lastmod timestamps, sitemap pings, and publishing volume don’t convince the scheduler. The two things that do are the two hard things: real popularity (links) and real quality improvements that show up in indexing signals and feed back into the scheduler. Everything else is theater. » (Tradução) (Síntese localizada do trecho quarenta e dois, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«## Google is trying to crawl less, not more » (Tradução) (Síntese localizada do trecho quarenta e três, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«The single freshest angle here, and one older guides miss entirely: Google’s own stated goal is to reduce total crawl volume, not grow it. In an April 2024 LinkedIn post, Illyes wrote: “My mission this year is to figure out how to crawl even less, and have fewer bytes on wire.” He pushed back on the idea that Google had slashed crawling — “we’re crawling roughly as much as before, however scheduling got more intelligent” — and framed the goal as a shared win: “Decreasing crawling without sacrificing crawl-quality would benefit everyone.” The mechanisms he pointed at were better caching, cache sharing across user-agents, and fewer bytes transferred — not “crawl my site more.” » (Tradução) (Síntese localizada do trecho quarenta e quatro, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«The takeaway: crawl demand was never something to maximize. Google is actively optimizing for less total crawling with equal or better crawl quality, routing the demand it does have toward URLs more likely to deserve it. Your goal isn’t more demand — it’s correct prioritization of the demand you’ve earned. » (Tradução) (Síntese localizada do trecho quarenta e cinco, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«## Do you even have a demand problem? » (Tradução) (Síntese localizada do trecho quarenta e seis, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«Most sites don’t, and shouldn’t spend a minute on this. Google’s own de-escalation applies squarely to demand: “If your site doesn’t have a large number of pages that change rapidly, or if your pages seem to be crawled the same day that they are published, you don’t need to read this guide.” John Mueller has been similarly blunt about scale — per Search Engine Roundtable’s coverage of his tweet, 100k URLs is usually not enough to affect crawl budget, since that works out to well under one crawl per minute over three months. » (Tradução) (Síntese localizada do trecho quarenta e sete, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«If you are big enough to care, here’s the diagnostic that separates a demand problem from a capacity one — with a caveat up front: it produces a hypothesis to test, not a diagnosis. Pull up GSC’s Crawl Stats report and your server logs. If host status is healthy and average response time is fine but a set of URLs is barely getting crawled — and they’re stuck in “Discovered – currently not indexed” — that pattern is evidence pointing toward demand, not proof of it. Crawl Stats shows crawl activity (a capacity-side view), not a demand score — there is no public per-site “crawl demand score” anywhere, so you’re always inferring demand from activity plus indexing status, never reading it off a dashboard. » (Tradução) (Síntese localizada do trecho quarenta e oito, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«Before you act on “it’s a demand problem,” rule out the other things that produce the same healthy-server-but-not-crawled symptom: Google may not have discovered the URLs yet (no path in, no sitemap entry), rendering may be hiding the content Googlebot needs to see, canonicalization may point Google somewhere else entirely, real quality issues (thin, duplicate, low-value) can get a page crawled but deliberately left out of the index, and Google’s own indexing selection can sit a page out even when crawling and quality are both fine. Only after those are checked and don’t explain it does “low demand” become the working explanation — and even then, treat it as the best-supported hypothesis, not a confirmed cause. Faster hardware won’t fix a genuine demand problem; the fix is on the demand inputs: links to those pages, genuine reasons to recrawl them, and less junk inventory drowning them out. » (Tradução) (Síntese localizada do trecho quarenta e nove, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«For ground-truth, per-URL crawl data, log analysis is the answer. I’ll flag one current tool I can speak to firsthand in the Tools tab. » (Tradução) (Síntese localizada do trecho cinquenta, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«## Crawl demand vs. rate vs. budget vs. frequency » (Tradução) (Síntese localizada do trecho cinquenta e um, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«Keep the family straight: » (Tradução) (Síntese localizada do trecho cinquenta e dois, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«- Crawl demand — how much Google wants to crawl (popularity + staleness + perceived inventory). This page.

  • Crawl rate — how fast it can (capacity / host load). Its own page.
  • Crawl budget — the two together: “the number of URLs Googlebot can and wants to crawl.”
  • Crawl frequency — how often a given URL gets recrawled, which is a demand output (mostly staleness).
Evidência desta afirmação Google's crawl-demand guidance discusses perceived inventory, popularity, and staleness. Escopo: These inputs influence crawling but do not provide a user-controlled demand setting. Confiança: alta · Verificado: Google: Large site crawl budget guide

» (Tradução) (Síntese localizada do trecho cinquenta e três, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

«Bing doesn’t use the term “crawl demand” — it reframes the whole thing as crawl efficiency: Fabrice Canel defines it as “how often we crawl and discover new and fresh content per page crawled,” and Bing’s philosophy is inventory-reduction-first, the demand-side equivalent of perceived inventory. IndexNow is Bing’s way of signaling demand-relevant change events instead of waiting for the scheduler to infer staleness — but note Google does not use IndexNow, so it won’t move Google’s demand.

Evidência desta afirmação Google's crawl-demand guidance discusses perceived inventory, popularity, and staleness. Escopo: These inputs influence crawling but do not provide a user-controlled demand setting. Confiança: alta · Verificado: Google: Large site crawl budget guide

» (Tradução) (Síntese localizada do trecho cinquenta e quatro, parte um: o texto-fonte foi preservado para conferência na revisão nativa.)

Adicionar uma nota de especialista

Fixar uma citação de especialista

É uma pessoa nova? Crie o perfil não reivindicado dela em /admin/experts/ → Fixar uma citação de especialista primeiro.