Few hosting decisions attract as much bad advice as the shared versus dedicated IP question. Sales pages sell the upgrade as a ranking lever. Agencies parrot it. And site owners end up paying a monthly premium for something nobody can actually measure. The technical reality is narrower, and honestly more interesting: your address model shapes deliverability, security reputation and resource contention, and those touch search performance sideways. Figure out where the causal chain really runs and you save money - plus you stop misdiagnosing every traffic dip.
How Search Engines Actually Treat IP Addresses
Every crawl starts with a DNS resolution. Hostname becomes an address, then the fetch happens. So yes, the address sits in the request path. No, it is not a documented part of the ranking formula. Search engineers have said this over and over: a site gains no inherent strength from sitting on its own address. Nothing in the published guidance contradicts them.
What the address does give you is infrastructure context. It helps identify hosting neighbourhoods, feeds spam detection heuristics, informs crawl scheduling. None of that is a quality score. And the correlation people keep pointing at - sites on dedicated addresses tend to rank well - just reflects who buys dedicated hosting. Larger organisations. Better funded. Maintained codebases, real editorial output. Causation runs from the budget, not the octets. Whichever you pick, run the checks before buying a range.
Genuine weight shows up in three places: geolocation hints used for country targeting, reverse DNS verification, and reputation systems run by mail providers and security vendors. Those matter. But none of them is a ranking factor in the way the sales pitch implies.
Shared Hosting: The Neighbourhood Problem, Measured Honestly
On commodity shared plans, hundreds or even thousands of hostnames can answer on a single address, which bundles a pile of unrelated businesses into one reputation record. Sounds alarming. Then you look at how penalties actually propagate: search engines assess pages and hosts individually, so a neighbour caught spamming does not automatically drag down everyone else on the machine.
Risk concentrates at the extremes. An address packed with link networks, malware distribution and bulk doorway domains is a completely different animal from one hosting local plumbers and dentists. And the sharper danger is not ranking suppression - it is blocklisting. Security scanners and browser warning systems work at address level, so a neighbour’s malware can throw interstitial warnings that gut your click-through while your positions sit there looking fine.
- Reverse IP lookup to see who else answers on your address
- Blocklist queries against the major security and mail reputation databases
- Search Console security issues report, checked on a schedule rather than after a crisis
- Browser warning checks from a clean network, not a cached local session
Tip: run a reverse IP lookup before signing a shared contract, then repeat it quarterly.
Performance, Core Web Vitals and Crawl Budget
Server response time is where hosting genuinely meets search. It feeds straight into Largest Contentful Paint and shapes how aggressively crawlers schedule requests. On oversubscribed shared plans the bottleneck is rarely the address - it is CPU throttling, exhausted PHP workers and I/O contention between tenants fighting over the same spindles or NVMe queue.
Slow or erroring responses teach crawlers to back off, and that delays discovery of new and updated pages. On a small brochure site? Costs you nothing measurable. On a publisher pushing dozens of URLs a day it compounds into stale indexes and missed news cycles. Dedicated resources help by removing noisy neighbours from the equation, though I have seen a well provisioned container beat an underspecified dedicated box more times than the vendors would like.
Diagnostics worth your time: Time to First Byte sampled across a full week, Crawl Stats report trends, server error rates by status code, and PHP worker saturation during traffic peaks.
Tip: measure TTFB at peak hours, not once at midnight when the machine sits idle.
SSL, HTTP/2 and the Technical Case That Expired
The classic justification for a dedicated address was certificate installation. Server Name Indication killed that argument for every browser still in circulation, and modern stacks issue and renew certificates automatically no matter which address model you run. Anyone still selling dedicated IPs on SSL grounds is reading from a decade-old manual.
There is even a case where sharing helps you. HTTP/2 connection coalescing lets a browser reuse one connection across hostnames that share an address and certificate, which trims handshake overhead.
- Direct IP access for pre-DNS testing and staging verification
- Firewall allowlisting where a partner or payment provider whitelists your origin
- Custom port services that cannot be virtual-hosted by name
- Transactional mail sending, where reputation is bound to the sending address
- Legacy compliance frameworks that specify isolated infrastructure
Outside that list? The upgrade is a product being sold, not a problem being solved.
International Targeting and Server Location
Address geolocation is a weak hint, and it gets overridden easily by country code domains and explicit Search Console targeting. A German site on a Dutch address confuses nobody once the stronger signals are in place. Physical distance still matters for latency, sure, but latency is a user experience factor rather than a direct ranking input.
A content delivery network decouples delivery location from origin location and usually solves the latency question far more cheaply than relocating your hosting ever will. For multi-market operations the effort belongs somewhere else entirely: correct hreflang annotations, local currency and payment methods, content genuinely adapted rather than machine-swapped. Address geography sits well below all of that.
Tip: verify how third party databases geolocate your address, since stale records can misplace a server by continents and distort your own analytics before they distort anything else.
When a Dedicated IP Is Worth the Line Item
Build the decision on four variables: traffic volume, transactional email volume, compliance exposure and how much shared reputation risk you can stomach. Ecommerce and lead generation operations pushing big mail volumes benefit most, because sending reputation attaches to the address - which means a neighbour’s bulk campaign quietly becomes your inbox placement problem. Publishers with heavy crawl demand benefit from resource isolation, not from the address as such. Small brochure sites rarely register any change at all, and that budget does more good spent on caching, content or link acquisition.
- What resource guarantees are contractual rather than aspirational?
- What is the address history, and who used it previously?
- How does the abuse policy work, and how fast do you act on reports?
- What downtime does migration involve, and what is the rollback path?
Tip: ask explicitly for address history and confirm it was not previously operated by a spam sender.
Migration Without Losing Visibility
A clean move follows a predictable sequence. Lower DNS TTL well in advance, stage a full copy on the new node, verify it through host file overrides, switch DNS, then keep the old node serving until propagation finishes. Skip the verification step and that is where your incident comes from. Every time.
The recurring failure modes are mundane, not exotic: mismatched PHP versions breaking plugins, redirect rules abandoned in the old configuration, staging robots files that block crawlers after go-live, certificates that expire at cutover because renewal was tied to the previous host.
For the fortnight afterwards, watch server errors, Crawl Stats, index coverage, TTFB and crawler behaviour in the raw logs.
Tip: archive server logs from before the migration so you can compare crawler behaviour instead of guessing.
Tip: never migrate hosting and change URLs in the same week, or you forfeit the ability to attribute any drop.
The Verdict for Practitioners
The address model is an infrastructure decision with indirect search consequences. Not a ranking lever. What genuinely moves positions on the hosting side is deeply unglamorous: stable uptime, fast responses, clean security status, and resource allocation that matches what was actually sold to you. The address is only part of the picture, because of the footprints SSL and a CDN add.
Buy a dedicated address when mail reputation, compliance or operational control justify it, and treat any ranking improvement as a side effect of better performance rather than of the purchase. When a vendor promises ranking gains from an address upgrade, ask instead for CPU allocation, memory limits, worker counts and documented response time targets. Those numbers predict search outcomes. The address does not.
