I like shopping. Not that I particularly enjoy going to stores, or the acquisition of goods, or spending money, but for me shopping is really about research. And I love research.

When I discover (or decide) I need some new thing, I’ll immediately and excitedly block off a few hours for my upcoming study session. Which are the best ¹“Best” meaning here a healthy mix of objective superiority and subjective preference. versions of that thing? What brands are reliable? Are knockoffs good value, or compromised imitations? Where can I get the most value for my Canadian dollar? And, most importantly, what are the core attributes I should check for when browsing options?

To take a recent example, I recently decided that the green coating of pollen covering our windows had to go and bought myself a pressure washer. These days, unless you have to do some serious industrial work, the best pressure washer is going to be electric (lower maintenance, more reliable, lighter, no need for gas, albeit less powerful), and the best of those is going to have a brushless motor. There are a few big-name brands, but knockoffs are surprisingly acceptable given that even the most popular models have reliability issues, so the delta between mainstream quality and off-brand isn’t necessarily that large (though neither is the cost difference). That said, you should still make sure that you get one with metal fixtures (most common point of failure), and that hits a minimum of about 1800psi and 1.2 gallons per minute. Take those criteria, match it against your personal preference, budget, and preference for certain bells and whistles, and it’s easy to find a model that’ll be a quality fit for your needs.

There are a few notable kinks to this process – particularly in categories where there simply are no good options, like printers. In those cases I’ll waste a few quiet hours of my life, getting gradually more desperate as I slowly disqualify every major product line and eventually settle for something that looks pretty (because if it’s going to end up a paperweight, might as well be a pretty one, right?). But, for the most part, it serves me well and brings a great deal of enjoyment to my life.

Until a light burns out. Every time I see a bulb flicker my stomach drops, my anxiety rises, and I consign myself to a few more days circling the drain of lightbulb shopping. What’s unique about lightbulbs? Well, the problem is that, unlike printers, there are good options. There’s a surprising range of quantifiable aspects of lightbulbs that make a given bulb “good,” but actually finding one to buy that meets all those success metrics is an endless cycle of frustration. So what makes it so challenging? Hey, if you’ve made it this far, I’m assuming you’re interested, so let’s go down the rabbit hole of quantifying lightbulbs.

What makes a good lightbulb?

Warmth

If you’ve shopped for lightbulbs before, this is probably the first thing that came to mind. Lights can be “warm,” which is to say that they have a yellow/orange cast or “cool,” in which case they have a much bluer tint. Warmer lights (lower on the scale) are cozier and softer, but too warm and things start to look dirty. Colder lights (higher, starting around 4000k) have more of a crisp, clean, bright look, but can be harsh and antiseptic as you get too far down the scale. Picking the right warmth for the right room (warm for bedroom, cool for bathroom, etc.,) is important, but I find my all-purpose warmth is about 3500 kelvin. 3000 kelvin is the most popular option but verges on a bit too warm and muddy for my taste (and don’t give me any of that 2800k nonsense), whereas by 4000k things are starting to look a little harsh. Thankfully, 3500 kelvin isn’t overly uncommon, so this isn’t the most significant constraint.

Brightness

Not too much needs to be said about brightness, but it’s still necessary to consider to weed out the real oddball bulbs. Brightness is measured in a unit called “lumens,” and the sweet spot for general purpose lighting is about 800 lumens. This is about equivalent to a 60 watt incandescent bulb, for those of you old enough to remember when those were the only option. Any high end use cases will usually call for closer to 1600 lumens, and a moody side light might be closer to 400.

CRI Rating

CRI rating is where things start getting tricky. CRI stands for Colour Reproduction Index and, put simply, is a measure of how closely the light output by a bulb will reproduce colours as they would be seen under natural sunlight. Because the light emitted by a light source is not produced in uniform intensity across the visible spectrum, a light source with an excessively skewed output will mute or oversaturate certain colours, dependent on the distribution.

Sunlight, with a relatively uniform distribution, is treated as the norm, and so we measure lights on their ability to reproduce a key set of 15 colours, scoring each on a scale of 1-100, and taking the average to get the aggregate CRI scores.

A light with an average CRI score of 90+ is generally seen as having good performance, while top performers are rated 95+. It’s notable that there are a number of complaints with how manufacturers grade their performance against this scale (and with the system itself), but broadly a 95+ CRI rating will mean that the colours in your home (both paintings and people) will look bright and lively. Except…

R9 Rating

Given that we have this convenient measure of colour reproduction across 15 different swatches, you’d think we’d be taking the average of the full set, right? Hah, nice idea, but no. Turns out the CRI that is advertised on most bulbs is the “General CRI” (sometimes referred to as Ra or CRIa), which only takes the average of the the first eight colours. So even if a light performs abysmally on the next seven, it could still get a 95+ CRI score by specifically optimizing for the first set. Why is this the standard? I honestly have no idea, but I’m chalking it up to the deep pockets of Big Lightbulb.

This is particularly relevant because of the recent technological shifts in bulb making. I’ll review the options a little later on, but suffice to say that the most popular bulb type these days is LED, and it just so happens that it’s extremely common for LED bulbs to perform terribly on the ninth colour of the CRI swatch, generally referred to as R9. So those high scoring LEDs that you just rushed out to buy? They’re going to make anything with red tones look awful (which, notably, includes some skin tones, a major drawback to LED lighting). Unfortunately, most manufacturers don’t advertise their R9 scores anywhere, outside of a few specialty manufacturers. Thankfully there are, as always, intrepid internet researchers who have spent a great deal of money on specialty gear to measure the R9 scores of popular bulbs as well as the CRIe score, which is short for CRI Extended and represents the average of the first 14 swatches ²TCS 15 is was added much later to the core set, hence CRAe being generally measured as just the first 14. The 15th was added by the Japanese Industrial Standards organization (see JIS C 1609-1:2006 Illuminance Meters).

Shape and Socket

Of course, whether or not a light is good or not is irrelevant if it doesn’t fit the socket or the space. Most of the time hopefully you can just buy A19s and spend no more time thinking about bulbs than is strictly necessary (‘A’ lights are the familiar bulb shape with a standard light-socket fitting, 19 is a measurement of size and is the most common measurement). But, if you’re lucky like me, you may have no less than six different bulbs to shop for, each of varying levels of obscurity. For the most part my house calls for BR30s or, in common parlance, wide-ish flood lights. But I also have to accommodate T3s, GU10s, B10s (with their cute little base), E11s, and of course the venerable A19s.

Cost

In some cases, when there’s a mire of mediocre options in a product category, you can solve most of your problems and get something much better by paying just marginally more than average. Pens are a good example of this – pens you might pick up at the grocery store frequently have cloggy ink, write scratchily, and have only a passing understanding of ergonomics. But, for just a few dollars more you can get a four pack of Uni-ball Vision Elite Roller pens and have a much better time writing for years to come. Unfortunately, when it comes to lightbulbs, that’s very much not the case. Not only is there a pretty significant variance in price among even the more mediocre entries, but the price for premium products goes up extremely quickly. As an example, the truly excellent 95+ CRI BR30 bulb from San Francisco’s Waveform Lighting goes for $28 US dollars, and that’s certainly not the top of the line.

There’s not much I can offer in terms of in advice in this particular dimension, of course everyone’s ideal price point is going to be different, but unless you have an absurdly high budget for lighting (or very few sockets), this isn’t going to be a problem that a little more money will take care of.

Other things to consider

Of course like any other product, there are many more qualities that differentiate specific entries in the category, but that are not of sufficient importance to affect my decision making outside of a head-to-head between otherwise ideal options. That said, these are still worth enumerating, as they may be more important in certain contexts or could more strongly affect your preferences than mine. Directionality and uniform diffusement may matter more or less, depending on your specific need, as may dimmability (and make sure you have a compatible dimmer, or they will buzz and flicker like there’s no tomorrow!). Some might argue that the lighting technology is a matter of choice (incandescent, fluorescent, LED), but in all earnestness LED is the only reasonable choice – longer lasting, more energy efficient, and better for the environment than any of the alternatives.

The one quality which I haven’t included as a primary deciding factor, but may be a point of contention for the more discerning lightbulb enthusiasts, is lightbulb flicker. Depending on how an LED cycles, it’ll have different presentations of frequency, modulation, and ‘duty cycle’, but in my experience no bulbs that I’ve bought of reasonable quality have had visible flicker issues. As such I don’t give it particular consideration, but if you are particularly sensitive to that sort of thing, it’s definitely worth taking into account.

Was that really so bad?

So it’s simple then, right? Just find a listing for a 3500k, 800 lumen BR30 lightbulb, with a CRIa score of 90+, a good R9/CRIe score as calculated by some strangers on a forum, and that costs less than $10/bulb. Well, turns out that even if you do manage that and buy a dozen, if you go back six months later that particular bulb is likely to be gone, replaced by a marginally but critically different model that is somehow lacking or at least undocumented. There exists a mind-boggling array of bulbs from any given vendor, and the turnover is bizarrely rapid, making any precise model extremely challenging to find from any larger company. Moreover, and this is likely relevant to less of my audience (pretending that any of this was relevant, and that I have an audience), but it insanely hard to find lightbulbs in Canada. Many major brands, like Cree and and Waveform are near impossible to find (especially for specific high-performing models), and others are many multiples more expensive than their US counterparts. Many, many times I’ve finally found an high-quality, appropriately priced bulb that seems common enough that surely someone in Canada must carry it, only to be stymied time and again.

So what’s the takeaway here? Well, at the least I hope you learned something about lightbulbs, but more-so this has been a selfish act of catharsis for me, putting to words something that’s been bothering me for years. In the past decade I’ve lost nearly a week of my life to googling lightbulbs and that’s not something I’m proud of. Perhaps I could just get less-great bulbs, but now I know enough to be bothered by their small underperformances. The healthiest option is likely to never buy lightbulbs again, to hand it off to someone else who can not stress the details quite so much, but same problems goes. So I guess the real takeaway is that it’s too late for me, but maybe you can use this to save yourself some trouble while still getting a good product, next time you have to buy a lightbulb yourself.