The venerable Paul Graham postulated in a blog post recently:

“If you have a US startup called X and you don’t have x.com, you should probably change your name.”

This is an oversimplification. Whether a domain ends in .com is just one of several important characteristics to consider in naming a startup. When (and when not to) splurge on a .com is the more fundamental question (hint: skip to the end for the answer, or stick around for my story).

If you’ve ever named a company, I’m sure you can testify that it takes research, imagination, and an open mind. I recently went on that journey myself, and learned several things along the way: like some of the criteria to consider, how to use those criteria to effectively evaluate names, and when it’s worth going .com. Here’s how the story unfolded for me.

Naming Crema.co

Around this time last year I was in the early process of conceptualizing and naming Crema.co, a coffee marketplace with a Netflix queue-like subscription model. I’ve been complicit in choosing names for a few companies over the years, but this was my first time naming an online consumer brand. Not only did the name itself need to be meaningful, but the domain name had to be short, memorable, and affordable.

As a result, I spent months trying to come up with the perfect name. I had brainstormed well over 300 (mostly bad) ideas, made around 100 enquiries with existing domain name owners, and had a “shortlist” 50 entries long. Every car trip I coerced my wife into helping me come up with ideas. I accosted friends and family to run names by then. In a word: I was obsessed.

The first names that I thought of were boringly straightforward. Because of this, they were also unaffordable (everyone else had thought of them too). The subsequent waves of names relied on misspellings, made-up words, and non-dot-com extensions to try to arrive at something attainable. This is where the real creative challenge lies. (Daniel Eckler’s post, How to Name a Unicorn, shares some great tips on the art of brainstorming.)

For me, there was a name that led the pack at any given time. Some of the leaders held their position for weeks. But uncertainty would creep back in, and the search would resume.

What’s in a name?

The turning point in my quest came when I began writing down the characteristics I was looking for in a name. It sounds obvious, but while I had inklings of these qualities from the beginning, putting them down on paper helped me to evaluate each name more objectively, and helped circumvent emotional attachment to weaker names. For me, the criteria clustered around three categories:

1. Meaning
2. Memorability
3. Cost

Meaning tends to be what drives ideation, while memorability and cost are funnels that filter out most of the possibilities that come out of brainstorming.

Within each of these categories, I articulated specific qualities that I was looking for.

Meaning

I needed a name synonymous with coffee that would connote hand-crafted quality, and perhaps also be extendable to other food and beverage categories. Specifically:

• On encountering the name for the first time, would someone have a clue that it’s a marketplace for buying coffee?
• On encountering the name for the first time, would someone have a clue that it is an online marketplace for other foods and beverages?
• Does it convey coffee/food/beverages hand-produced by expert craftsmen/craftswomen?

Memorability

Because I was building an online consumer marketplace, I needed a name that was short and phonetic; something people could easily remember and spell.

• Would you know how to spell it if you first heard it in conversation?
• Is it short in length (<5 characters), medium length (6-7 characters), or long (8+ characters)?
• Would you remember the name (including its spelling) if you had seen it once before?
• Is it hip and offbeat rather than boringly direct?

Cost

And finally, I needed a name that I could afford. This is where .coms are problematic. My bootstrapped budget for a domain was $10,000 (and preferably much less). While it would obviously be great if I could find a strong .com for that amount, I was happy enough if the equivalent .com was attainable for $50,000 or less; I could always buy it later once the company takes off.

• Is it doable ($5-10k), affordable ($1-4.9k), or virtually free (<$1k)?
• Is it a .com? If not, could the .com be attainable later for 50k or less?

Evaluate

Once I had formalized these criteria, I ended up throwing them into a spreadsheet so I could compare my name candidates side by side. To keep things simple, I decided to give each name a thumbs down (0), neutral (1), or thumbs up (2) for each criterion. I then tallied up the subtotals for meaning, memorability, cost (points awarded / total possible points), and averaged the three subtotals together to arrive at the final score, expressed as a percentage.

At this point it’s important to keep in mind that anytime a quantitative system is used to evaluate subjective qualities, the results should be interpreted with several grains of salt. This model assumes I’ve identified all the right criteria, evaluated each name objectively, and weighted the scores proportionally, none of which are completely true. In leu of perfection, the smell test will have to suffice.

In the end, Crema.co scored the highest of the 10 names I evaluated using the spreadsheet, with a total score of 81%. Before doing the analysis, I had favored Crave.co and Stache.co, which both scored around 50% by comparison. And earlier I’d been irrationally attached to Yarn.co, which garnered a dire 21% in the analysis. Had I not gone through this exercise, I may have opted for an inferior name.

To .com, or not to .com?

So, back to the opening question: should startups opt for a .com, as Paul Graham suggests?

Graham rightly points out that having a .com signals strength. This is undeniably true; almost all highly successful online companies now have a short .com domain name.

The key word being now. As John Doherty recently wrote:

“What Graham failed to mention is that 4 of the top 8 [YCombinator companies] started with a different domain name than what they ended up with.”

Dropbox started with getdropbox.com. Stripe was getstripe.com. Airbnb was airbedandbreakfast.com. Facebook started as thefacebook.com. And more recently, plenty of other companies have started with .co and .io domains.

The trick with .coms is that they’re expensive. It’s extremely unlikely that you’ll find a suitable .com that hasn’t already been registered, which means you have to buy it on the secondary market. And to put cost in context, the current owner of crema.com is asking six figures. (Morgan Linton’s post What Every Startup Founder Should Know About Buying Domain Names provides a healthy dose of realism.)

The 1% rule

The real question is not, “Should you go .com?”, but rather, “How much should you spend to get a strong domain?”

I’d like to suggest a rule of thumb:

Don’t spend more than 1% of your company’s valuation on a domain.

Let’s say you’re a new startup and you’ve raised $100,000 in seed funding at a $1,000,000 valuation. It would be foolish to blow half of your capital on a $50,000 domain name. At that point in the company’s life, whether the name ends in .com will have almost no impact on whether or not you’ll make it to the next stage. Cutting your runway in half by waving goodbye to $50,000 (and 5% of your company), however, will have a trememdous impact. In this circumstance, $10,000 is the maximum you should consider spending — 1% of the company’s $1 million valuation.

If, on the other hand, you’ve built a successful company and just raised a $2 million round at a $10 million valuation to accelerate growth, it might be time to spring for that short and sweet .com for as much as $100,000 (again, no more than 1% of the company’s valuation).

Rules of thumb are, by definition, sweeping generalizations. A short .com deliveres less value to a B2B company than to a consumer brand, for example, so B2Bs should probably spend less.

The main point is to think about domain name purchases in terms of your current valuation.

Looking back

In naming Crema.co, the biggest thing I learned was the importance of defining the characteristics I was looking for in a name. Getting those out of my head, onto paper — and in my case, into a spreadsheet — helped me overcome irrational emotional attachments and think more objectively about each potential name. And, contrary to Paul Graham’s advice, I’m content with a non-dot-com. For now.