Choosing a TLD for Your Startup: A Decision Framework

Start with the constraint that cannot be reversed

You can change a website's hosting, logo or pricing in a day. Changing the domain your customers type, your email addresses, and every link ever pointing at you is slow and leaks traffic. So the first rule is to pick an extension you can live with for five years, and the second is to pick it before you print, ship or announce anything.

The framework below asks six questions in order. Answer them and a short list usually falls out without needing opinions about which TLD is "best".

Question 1: who is the buyer and what do they expect?

Buyers apply a familiarity test within a second. People outside the tech industry assume a business lives at .com; typing yourbrand.com by reflex is the default, and any other ending risks misdirected traffic to whoever owns the .com. Developers and technical buyers are comfortable with .io, .dev, .ai and .app, and in that audience an unusual ending can even signal belonging.

The test is simple: describe your product to ten target customers, show them your URL in text only, and ask where they would go. If a meaningful share would type the .com instead, plan to acquire it (see the end of this guide) or choose a name that is distinctive enough that confusion is harmless.

Question 2: will your email arrive?

Your domain is also your sender identity. Mailbox providers score sending domains on history, authentication and reputation, and the TLD is one input among many, not the deciding one. Practical consequences:

• Cheap, high-volume TLDs have attracted a lot of bulk abuse over the years, and some filters are stricter with them. Reputation varies by TLD and changes over time; blanket claims such as "TLD X is always blocked" are generally wrong, but test before you rely on it.
• A brand new domain has no sending history regardless of extension. Plan to warm it up gradually.
• Correct SPF, DKIM and DMARC records matter far more than the TLD. Set them before you send the first customer email.
• Before committing, send test messages from the new domain to several major providers and check placement, not just delivery.

If your business is outreach-heavy, use the most conventional TLD you can for the sending domain, even if the marketing site uses another.

Question 3: what does search actually do with the extension?

This is where most startup advice is wrong. Google's own documentation says the following about how it classifies TLDs, and it is worth reading in full: Google's guide to managing multi-regional sites. Its key points:

• A TLD that is not a country-code domain is treated as a generic TLD. Newer extensions such as .shop or .tech are in this group, so they have no ranking handicap compared with .com.
• Country-code TLDs act as a geographic signal. A .de site is read as targeting Germany, which helps for local intent and can hurt if you want a global audience.
• Google lists some ccTLDs it treats as generic because the world uses them that way. The list it publishes, which it says may change, currently includes .ad, .ai, .as, .bz, .cc, .cd, .co, .dj, .fm, .io, .la, .me, .ms, .nu, .sc, .sr, .su, .tv, .tk and .ws.

The myths to drop: that .com ranks better because of the extension, that "keyword TLDs" like .seo boost rankings, and that a spammy neighbour TLD drags you down automatically. Search quality follows content, links and brand demand. Brand recall is where the TLD does matter, because people who misremember your ending search for you less.

Question 4: do you want the browser to force HTTPS for you?

Some TLD operators have added their entire extension to the browser HSTS preload list, which means browsers connect over HTTPS only for every domain under it, even on first visit. Google Registry has done this for .app and .dev. That means a site there must serve a valid certificate from day one, or it will not load. The upside is a built-in guarantee for visitors and an implicit security signal; the cost is that a misconfigured staging site breaks for everyone. The preload list site hstspreload.org explains how TLD-wide preloading works and lets operators and domain owners check status. If you plan to use a plain-HTTP internal tool on that domain, choose another extension.

Question 5: what will it cost in year three?

Domain prices are rarely what they look like on the front page. The pricing traps in order of frequency:

1. Promotional first year. A very low first-year price followed by a renewal at a multiple of it. Always read the renewal price before checkout.
2. Premium renewals. Registry-designated premium names can cost more every year, not just at purchase.
3. Registry price increases. Some contracts allow periodic increases, and some registries have changed prices notably. A TLD that is cheap today is not guaranteed to be cheap in three years.
4. Add-ons. Privacy protection, email and DNS bundled at checkout, and transfer fees when you leave.
5. Multi-year lock. Registering for several years up front protects against price changes at the cost of cash.

Compare TLDs by the total cost over your planning horizon, not by the headline number. The budget new gTLDs compared guide walks through the cheap end of the market and the behaviours to expect.

Question 6: is the name clean, legally and reputationally?

• Trademark. Search national trademark registers and the common-law use of the name before you buy a domain. Owning a domain does not give you rights in a name, and a trademark owner can use dispute procedures such as the UDRP against a domain they consider infringing. If you are new, a domain that merely resembles an established brand is a risk across every TLD.
• Neighbour check. Look up whether the same label is registered in other TLDs, and by whom. A competitor holding your name in .com is useful information, not just an obstacle.
• History. If the name was registered before, check its past for spam and blocklist listings. The evaluate an expired domain checklist applies to any second-hand name.
• Look-alikes. Decide which close variants you will defensively register. The cost is small next to dealing with impersonation, as discussed in typosquatting and look-alike domains.

Decision table by startup type

If your startup is...	Reasonable default	Why, and what to watch
Consumer product, mass market	.com	Highest familiarity; avoid anything the customer has to be told about
Developer tool or API	.dev, .io or .com	Audience is comfortable with these; .dev enforces HTTPS
AI product	.ai or .com	.ai signals the category and is treated as generic by Google, but check renewal and registry rules
Mobile or web app	.app or .com	.app enforces HTTPS; the name must be distinctive
Local service in one country	The country's ccTLD	Local trust and geotargeting; check residency rules
Online shop	.com, then a shop-themed extension for campaigns	Do not make a cheap extension your only domain
Nonprofit or community	.org	Established convention for the category
Pre-launch experiment	Any inexpensive TLD, planning a migration	Keep email off the experimental domain

For deeper head-to-heads see .com vs .io vs .ai vs .app and .dev vs .app. The reasons these labels behave differently, such as registry type and governance, are explained in gTLD vs ccTLD vs new gTLD.

Starting on another TLD and acquiring the .com later

Many companies launch on a different extension and buy the .com after funding. This can work, with discipline:

• Find out who owns it. Look up the registration data and whether the name is parked, in use, or near expiry. If it is in the late lifecycle stages, the backorder process may be an option; if it is held by a person, a broker or direct contact is the route.
• Use a broker or escrow. Contacting the owner yourself reveals your interest and can raise the price. An intermediary and an escrow service for the transfer are standard.
• Set a ceiling in advance. Value the name against the cost of the alternative: brand confusion, lost direct traffic and a later rebrand.
• Keep the old domain. After the move, 301-redirect the old domain permanently and keep renewing it, since existing links and customer habits persist for years.

The risk of the plan is that your own traffic builds value for whoever owns the .com. If the name you want is a distinctive coined word with no .com owner, buy it now; if the .com is occupied by an unrelated active business, stop treating "later" as a plan.

Frequently asked questions

Does Google rank .com higher than new TLDs such as .shop or .tech?

No. Google's documentation says TLDs that are not country codes are treated as generic, so new gTLDs get no ranking penalty or boost for the extension. Content, links and brand demand decide.

Will .io or .ai be geotargeted to a country?

Google lists .io and .ai among the ccTLDs it treats as generic, so they are not geotargeted to their home territories. The list can change, so check Google's current documentation.

Which extensions force HTTPS in the browser?

Google Registry preloaded HSTS for the whole of .app and .dev, so browsers require HTTPS for every domain there. Check hstspreload.org for current status of any extension.

Can a TLD hurt my email deliverability?

It can be a minor factor, since filters distrust extensions with a history of bulk abuse. Authentication (SPF, DKIM, DMARC) and sending reputation matter far more. Test inbox placement before committing.

Should I buy the .com if someone else owns it?

If it is a distinctive coined name and your audience is mainstream, usually yes, using a broker and escrow. If an unrelated business actively uses it, choose a different name instead of a long-term workaround.

How do I avoid renewal price shocks?

Read the renewal price at checkout, not the first-year price. Check whether the name is premium, consider multi-year registration, and compare total cost over three to five years.