Decluttering the Art Supply Closet, Part III: Choosing Colors for My Palette

Photo by Steve Johnson on Unsplash

Last time, I talked about cycles of expansiveness and restriction when it comes to art supplies and hobbies generally. Now, I’d like to apply it to palette colors.

How Many Colors?

Finding the “just right” number of colors is a balancing act. Having too few colors is a struggle because you’re constantly mixing, and it can be tough to get the colors you want; but having too many is also a problem, because there’s just too many factors to learn and keep in mind about each shade, and too much day-to-day decision-making between similar shades. If I owned all 240 colors in Daniel Smith’s line, I’d be like the Math Lady meme, constantly trying to figure out the perfect color to use in every situation taking into account mixing, granulation, opacity, etc. 

Math Lady meme. “Should I use Indian Red or Lunar Red Rock…?!”

I think the ideal is having a reasonable number of colors in a wide variety of hues and other characteristics, and knowing them well. Probably the ideal way to go about it is to start small and slowly add one additional color at a time, but since that’s not how I roll, I instead started by trying every color in every dot card I could find, and I’m eliminating options from there.

So how many is a “reasonable number?” As with all things in art, I think it depends on who you are and what you prefer. Does it bring you joy to have a lot of options, or to get a lot of mileage out of just a few options? Would you like to plan on typically having a rotating cast because you get bored easily, or would you prefer to choose your ride-or-dies and stick with them?

My answer is “yes to all,” so here’s how I manage it:

  • I usually try to pick just 3-6 colors for any given painting, but I also like to have a lot of options to “shop” from when I’m planning a painting.
  • I try to make sure all the core colors in my palette have a defined role, but…
  • I also leave room for rotating “guest stars”.

Color Choosing Criteria

In choosing colors, I used some of the same qualities I apply to supplies in general.

Basics: Is it expired?

Watercolor paint in tubes is generally expected to last about five years; pans about ten years. That doesn’t mean it can’t last longer, but after that time, be on guard. If it’s older than that and the pigment is separated from the binder; it’s moldy; it smells weird; or it’s too hard: that’s expired paint!

Joy: Does it spark joy?

Some colors just make me happy, even looking at them on the palette, or knowing that I can use them. If that’s true, it doesn’t matter how technically useful they are or are not. Similarly, other colors I simply dislike, and I know I will not use them no matter how many other artists insist they are must-have colors. 

Quality of Experience: Do I enjoy the process of using it?

I’m talking about the actual experiential feel of using a color, which is distinct from what color it is. Some colors are a joy to swatch out while others feel like a drag. I can’t always explain why. It has to do with tinting strength, paint texture, and how it responds to the amount of water and type of brush strokes you like to use…

Sometimes the same pigment is more or less enjoyable in different brand. Sometimes I just don’t like it no matter the brand.

If I don’t enjoy using a color, I will tend to avoid it, no matter how theoretically right it is for my palette.

Versatility: Does it have many different uses?

All colors have some use. I have been swayed to buy many colors I don’t actually use because of the sentence “This is the exact perfect color for [specific subject].” That sentence can be completed for any individual color, but I am rarely painting the exact subject I named.

Now I try to think beyond the most obvious “what color it is” and “what things are that color,” and consider the range of colors that can be mixed with this paint. Thinking about its mixes is usually a more reliable way to gauge its versatility, though I had to get a fair amount of practical experience mixing to be able to think theoretically about them.

Here are some rules of thumb for picking versatile mixing colors:

  • Primary colors are versatile (typically any shade of magenta, yellow, or cyan) because they are the building blocks of all other mixes. In general, I find a palette with multiple primary triads to be more versatile than one with lots of secondaries/earth tones. You can always mix secondaries and earth tones with reds, yellows, and blues, but not the other way around.
  • Bright colors are generally more versatile than muted colors. I was surprised to learn that when I began painting (because in home decorating and fashion, people often tell you that muted colors “go with anything” while brights are “garish”). But regardless of how you feel about muted vs. bright colors in your actual painting, bright colors have more capacity to mix various colors. You can always mix up muted shades using brights (by mixing complementary brights), but you cannot generate brights from muted paints. Whether you want to do the amount of mixing necessary to tame your brights is another question.
  • Single pigment paints can be more versatile than mixes, in the same way that buying eggs can be more versatile than buying cake. With Phthalo Green, you can mix up a blue green, a yellow green, etc. whereas a mixed color made with Phthalo Green and yellow may have more specific use case. Again, this comes down to balancing versatility and convenience. See also: Should I only get single pigment paints?

Specificity: How perfect is it for me?

The opposite of versatility, specificity is the quality of being perfect for the exact type of art or subject that I personally enjoy doing the most. I might not tell people in general to have this color, but for me, it’s the exact right shade for [subject I actually do paint all the time.]

Having a palette made up of primary triads was great when I didn’t know what I liked to paint or I painted “general subjects”, but once I started painting trees, suddenly greens and browns got a lot more interesting.

Uniqueness: Does it open up new options?

Paint companies offer many very similar shades. For example, I can’t really tell the difference between Hansa Yellow Deep (PY65) and Isoindolinone Yellow Deep (PY110), so I probably don’t need both of them on my palette.

If you are limiting yourself to X number of paints, you want to make sure they are all fairly different from each other and serve different roles in your palette, to widen the number of colors you are able to easily mix up.

It’s not necessarily a problem to have two very similar colors. If you really can’t tell the difference, it’s basically like you bought twice as much of a certain color of paint. You use up one then refill with the other, or you could put them on different physical palettes. I find it a little stressful to have extremely similar colors on the same palette; it causes me to have to make constant micro-decisions between those two colors. Do I use Quinacridone Rose (PV19), or Purple Magenta (PR122)? Is this more a situation for Ultramarine (PB29) or French Ultramarine (also PB29, but more granulating), or Ultramarine Green Shade (also PB29, but less granulating and a bit less violet-toned)? I’d rather “one and done” the decision by making it before I put the color on my palette.

Which colors you choose may depend on how large your palette is (and, er, how many palettes you have). In a small mixing palette, versatility is the name of the game, but in a larger (or more specifically-themed) palette, you may want to choose the perfect paint for a certain situation. For example, on my general travel palette, I tend to one earth orange (Transparent Red Oxide), but on my Desert Palette I have multiple earth reds/oranges (Quin Burnt Orange, Indian Red, Violet Iron Oxide).

One of the things I like to do with my palette is divide it into “slots,” e.g. “This is the color for my ‘primary magenta’ slot, this is the color for my ‘earth orange/burnt sienna’ slot, etc. etc.” (I talked about this in my post How to Build a Watercolor Palette from the Ground Up.) Slots help me make sure that all my bases are covered but I’m not reduplicating the same uses with multiple colors. If two colors fall into the same slot, I can remove one from my palette, or I can designate one as a “bonus color” (the way I do with Quin Coral because even though it is not my most useful red-orange, I just love it.)

Complementariness: How well does it go with my other paints?

In addition to figuring out if my palette bases are covered and if I have (at least/about) one of each “slot,” I also try to think about how my colors work together. Do they mix well?

As recommended in Nita Leland’s Exploring Color Workshop, it can be helpful to try to match paints with a similar tinting strength. It’s frustrating trying to mix delicate/weak colors with strong/overpowering colors. With other qualities (such as muted/bright or transparent/opaque), I feel there is value in having variety, but in tinting strength, I think it makes more sense to have a preference and stick with it for your palette.

In any palette larger than, say, 3-4 colors, it’s impossible to try to game out every single possible mix, but I like to keep in mind certain pairs I know mix well, or ensure that if I intentionally omit certain colors, I have the right components to mix a similar hue.

Here are some examples:

  • I used to use Quin Gold a lot, but I’ve been leaving it off more and more since I typically have the components to mix it: Nickel Azo Yellow and an earth orange/red such as Quin Burnt Orange, Transparent Red Oxide, or Quin Burnt Scarlet.
  • If I have a dedicated orange, like Pyrrol Orange (PO73), I may not feel that I need orangey-yellow (e.g. New Gamboge) or an orangey-red (e.g. Pyrrol Scarlet). I can mix orangey tones from my other yellows and reds. However, if I have/want New Gamboge and Pyrrol Scarlet anyway, I may feel that I don’t need an orange because I can mix it with the other two colors.
  • When choosing between two similar colors, like Phthalo Blue GS and Phthalo Blue RS, I like to keep in mind effects on other parts of my palette. RS may work alone for skies while GS may benefit more from being paired with a violet-blue such as Ultramarine. GS mixes brighter greens while the RS palette may benefit more from the addition of a green or turquoise for that purpose. The mixing complement of GS is more in the red direction, e.g. Pyrrol Scarlet, while the mixing complement of RS is more in the orange direction, e.g. Pyrrol Orange.


Choosing colors for a palette can be a tricky process with so many factors to keep in mind, but I think I also overcomplicate it because I enjoy it so much. It can also be as simple as grabbing a pre-selected pan set or just using your favorite colors.

I like to think through all these factors not only because it helps me feel that I’m optimizing my palette, but because considering justifications for adding/keeping/cutting colors can reveal emotional reactions I didn’t know I had. Kind of like tossing a coin to make a decision, and realizing from your reaction to it which choice you actually want to make.

If you think about cutting an “inessential” color and it makes you sad? Keep it! (Me and Quin Coral.)

But if you think about cutting an “essential” color and it makes you relieved? Do it! (Me and Perylene Maroon.)

If you think about adding an “essential” color and it makes you feel bored/put-upon? Leave it out! (Me and Raw Umber.)

But if you think about adding an “inessential” color and it makes you excited/energized? Do it! (Me and Prussian Blue.)

Even individual colors don’t do much for you emotionally, it can be really pleasing to use the palette-building slots to build a nice array that you know serves your needs for all your painting situations.