How to Plan Granny Square Colors & Layouts

A granny square blanket is decided twice: once when you choose the colours, and again when you decide where each square goes. This covers the second decision — how often each colour should appear, how to stop the same colour bunching together, how to number squares so the layout survives being unpinned, and how to work out how much of each colour you actually need.

Choose your colour palette

The palette is the one part of this that is genuinely a matter of taste, and nothing on this page will tell you which colours to pick. What it will do is point out the three practical constraints that decide whether a palette works as a blanket, as opposed to working as a set of colours.

Value matters more than hue. Value is how light or dark a colour is, independent of what colour it is. Two yarns can be very different hues — a mid rust and a mid sage — and still read as the same tone from across a room, which makes the boundary between squares disappear. The test is free: photograph your yarn together and look at the picture in black and white. Anything your phone's greyscale filter flattens into one shade will flatten in the finished blanket too. This is a check, not a rule; sometimes a low-contrast blanket is exactly what you want.

One colour usually has to give way. A palette where every colour is equally saturated tends to fight itself. Most workable granny square palettes have at least one quiet colour — a cream, a grey, an oatmeal — doing the job of separating the others. That is a common practitioner convention rather than a rule, and the equal model further down deliberately ignores it.

Count your dye lots before you count your squares. If a colour is carrying half the blanket you will need several balls of it, and balls from different dye lots of the same shade are not always the same shade. Buying the dominant colour all at once, from one lot, matters more than buying the accent colours that way.

Decide how often each colour should appear

Colour frequency is the number of squares that use a given colour. It is the single decision that most changes how a finished blanket reads, and it is independent of which colours you chose.

The useful way to think about it is that the grid fixes the total before any colour gets a look in. An 8 × 6 blanket is 48 squares, and that is true whether you use two colours or ten. So deciding frequency is really deciding how to cut 48 into parts — which means the parts have to be whole numbers, and they have to add up.

That has a consequence worth knowing before you start: the proportions you ask for and the proportions you get are rarely identical. Twenty per cent of 48 squares is 9.6 squares. You cannot crochet six tenths of a granny square, so something has to round, and the split you end up with is slightly different from the one you specified. This is not an error and there is no way to avoid it — it is what happens when a percentage meets a small whole number.

Equal versus weighted colour distribution

There are three shapes this decision usually takes. They are planning models, not recommendations: none of them is correct, and the differences between them are matters of intent rather than quality.

Equal gives every colour the same number of squares — 25 / 25 / 25 / 25 across four colours. It reads as a scrappy, busy, even-handed blanket where no colour is the subject. It is also the easiest to shop for, because you buy the same amount of everything.

Dominant palette gives one colour roughly half the blanket and lets the others punctuate it — 50 / 20 / 20 / 10. It reads as a blanket that is a colour, with accents. It is more forgiving of a palette where the accent colours are strong, because they appear rarely enough not to compete.

Exact inventory skips the percentages entirely. You count the squares you already have of each colour and plan around those numbers. This is how most scrap blankets actually get made, and it is covered in its own section below.

Three planning models on the same 8 × 6 blanket of 48 squares. These are planning examples, not recommended proportions.
ColourEqual
25 / 25 / 25 / 25		Dominant
50 / 20 / 20 / 10		Exact inventory
counted
A · Cream122418
B · Terracotta121012
C · Sage12911
D · Ochre1257
Total 48 48 48

The dominant column is where the rounding shows. Asking for 50 / 20 / 20 / 10 of 48 squares gives exact shares of 24, 9.6, 9.6 and 4.8. Rounding each down leaves you two squares short of the blanket, so those two go to the colours whose fractions came closest to the next whole number. The result is 24, 10, 9 and 5 — which is a delivered split of 50, 20.8, 18.8 and 10.4 per cent. Two colours were asked for the same share and did not get it, because one of them had to absorb the odd square.

Three eight-by-six grids of forty-eight granny squares, showing the same blanket planned three ways. Every square is labelled with its colour's letter and filled with that colour's own pattern, so the grids read without relying on colour. The equal model gives twelve squares of each of the four colours. The dominant model gives twenty-four, ten, nine and five. The exact inventory model gives eighteen, twelve, eleven and seven. All three total forty-eight because the grid decides the total. Twenty per cent of forty-eight is 9.6 squares, so the dominant model lands on ten and nine rather than 9.6 and 9.6.
The same 48-square blanket under each planning model. The total never moves; only the split does.

How to avoid accidental colour clumps

A clump is two or more squares of the same colour sharing a side. Whether it bothers you is a matter of taste, but it is worth knowing how often it happens by accident, because the answer surprises people.

Take the equal model above — 12 squares each of four colours on an 8 × 6 grid — and place them at random two hundred times. Every single one of those two hundred layouts contained clumps. The fewest was 4 touching pairs and the most was 22. Not one arrangement came out clean by luck.

That is the thing worth taking away: random does not mean evenly spread. Genuine randomness produces bunching, and a shuffled pile of squares will always give you some. If you want an even scatter you have to ask for one specifically, because chance will not supply it.

Avoiding clumps is a constraint you apply while placing, not a filter you apply afterwards. In practice that means: lay out the whole grid before you sew anything, place your most numerous colour first rather than last, and check each square against the one to its left and the one above it as you go. Leaving the scarcest colour until the end is what strands it, because by then the empty positions are wherever they happen to be.

Diagonal contact is usually left alone. On a grid, corners touch constantly, and a blanket where no two same-coloured squares meet even at a corner is very hard to arrange and not obviously better to look at.

Two eight-by-six grids of forty-eight granny squares, twelve of each of four colours in both. Every square shows its colour's letter and its own fill pattern. In the left grid the squares are placed with no rule about contact and twelve pairs of same-coloured squares share a side, each outlined. In the right grid the same counts are placed with contact ruled out, leaving no touching pairs. Across two hundred seeds an unconstrained layout of these counts produced between four and twenty-two touching pairs, and none came out clump-free by chance.
Same colours, same counts, same starting point. The only difference is whether contact was ruled out while placing.

Random versus planned layouts

“Random” and “planned” are not opposites here, and the useful distinction is not the one people usually draw.

A truly random layout is the pile-of-squares method: pull one out, sew it on, repeat. It is quick, it needs no preparation, and as the numbers above show it will produce clumps. Some people like the result. The genuine drawback is not aesthetic but practical — it cannot be undone. If you dislike how it is going at square thirty, the first twenty-nine are already joined.

A planned layout is decided before anything is sewn. That does not make it rigid; it makes it cheap to change. Moving two squares on a floor costs nothing, and moving two squares in a joined blanket costs an evening with a seam ripper.

The middle option is the one most people actually want: a layout generated by chance but constrained, so it looks unplanned without bunching, and then fixed in place before joining. That is what the granny square colour layout generator does — it places the squares for you under whatever contact rule you choose, and gives every layout a seed so you can return to the exact same arrangement later. Writing the seed down matters more than it sounds: blankets take weeks, and a layout you cannot get back is a layout you have lost.

How to create a granny square grid

The grid is rows multiplied by columns, and the arithmetic runs in either direction depending on what you already know.

If you know the square, you can find the blanket: a finished square of 6 in across, joined edge to edge, makes an 8 × 6 blanket of 48 in by 36 in before any border. If you know the blanket you want, divide by the square size instead. The granny square calculator does this properly, including the yarn side of it.

Two things are easy to forget at this stage. Joining adds width — a slip-stitch join adds almost nothing, while a join-as-you-go round or a woven strip between squares can add a quarter-inch or more per seam, which across six columns is an inch and a half of blanket you did not plan for. And the number of squares you need depends on the grid shape, not just the count: 48 squares will make an 8 × 6 blanket, a 6 × 8 one, a 4 × 12, or a 3 × 16, and those are very different objects.

It is worth settling the grid before the palette, because the grid is what turns a percentage into a number of squares.

How to number your squares before joining

This is the step most often skipped and most often regretted.

Once you have a layout you like, laid out on a floor or a bed, it exists in exactly one place. Somebody walks through it, a cat sits on it, or you simply need the room back, and the arrangement is gone. Colour alone will not rebuild it — if 24 of your squares are cream, knowing that position 8 is cream tells you nothing about which cream square goes there, and it will not distinguish it from position 20 either.

So number them. Work along each row: square 1 is the top left, square 6 ends the first row, square 7 begins the second, and the last square is 48. Write each number on a scrap of paper and pin it to the square, or use removable stitch markers with numbered tags. Then photograph the whole layout from directly above, with the numbers visible, before you touch anything.

The photograph is the backup for the numbering, and the numbering is the backup for the photograph. Either alone can fail; together they are very hard to lose.

Number in the same direction you intend to join. If you are joining in rows, number in rows; if you are joining in columns or in a spiral, number that way instead, so the next square is always the next number.

An eight-by-six grid of forty-eight granny squares laid out as a joining map, with rows labelled row 1 to row 8 down the left and columns col 1 to col 6 across the top. Each square is filled with its colour and its colour's own pattern and carries its joining number rather than its colour letter. Numbering starts at square 1 in the top left and runs along each row, so square 6 ends the first row, square 7 begins the second, and square 48 is the last. Each number should be written on a paper tag and pinned to its square, because colour alone will not identify which of the 24 cream squares belongs in position 8 rather than position 20.
Numbering runs along the rows. The map is what lets you take the layout off the floor without losing it.

How to plan a scrap-yarn granny square blanket

Scrap blankets invert the whole process. Normally you decide the proportions and then make the squares; here the squares already exist and the proportions are whatever they are.

Start by counting, by colour, what you actually have. Not by weight and not by eye — count the finished squares. That count is your palette and your frequency in one, and it is not negotiable in the way a percentage is.

Then find a grid that the total fits. This is the constraint that catches people out: your squares have to fill a rectangle exactly, so a total of 47 will not do, and neither will 49 unless you are prepared to make a 7 × 7. Useful totals have several factor pairs. If you have 47 squares, make one more and you have 48, which fits 8 × 6, 6 × 8, 4 × 12, 3 × 16 or 2 × 24. If you have 50, you are choosing between 5 × 10 and 2 × 25, and one of those is a scarf.

Where a scrap blanket usually goes wrong is a single colour with too few squares. One square of a strong colour in a blanket of 48 does not read as an accent; it reads as a mistake, and it draws the eye straight to it. Either make two or three more of it, or leave it out. That is a judgement, not a calculation, and it is one worth making before you join rather than after.

The generator's exact-count mode is built for this: type in the counts you actually have, and it will arrange them and tell you plainly if they do not add up to the grid.

How much yarn does each colour need?

There is no multiplier for this, and you should be suspicious of any figure that claims otherwise. How much yarn a granny square takes depends on the yarn weight, the hook, your tension and how many rounds the square has — a three-round square in DK and a six-round square in aran differ by a factor of several, so no published number can stand in for yours.

What works instead is one measurement and some arithmetic. Make one square, exactly as you intend to make the rest. Weigh it. Then multiply.

Yarn per colour for the dominant-palette layout, worked from one weighed square. The 11 g figure is this example's measurement, not a standard.
ColourSquaresYarn at 11 g per square
A · Cream24264 g
B · Terracotta10110 g
C · Sage999 g
D · Ochre555 g
Whole blanket 48 528 g

The worked example above uses 11 g per square, which is this example's measurement rather than a standard — yours will differ, and the point of weighing is to find out by how much. Once you have your own figure, the yarn for a colour is simply its number of squares multiplied by that weight, and the colour totals add back to the whole blanket as a check that nothing has gone astray.

Two things this arithmetic does not include: the yarn taken by joining and by any border, which for a join-as-you-go method can be significant, and the safety margin you should add for a dominant colour. Running out of the accent used five times is an inconvenience; running out of the colour used 24 times, in that dye lot, is a different problem. The yarn yardage calculator converts weights into balls and metres once you know your per-square figure.

Worked layout examples

All three examples use the same 8 × 6 grid of 48 squares and the same four colours, so the only thing changing is frequency.

Equal. 12 squares each. Every colour appears the same number of times, no colour leads, and the blanket reads as scrappy and even. Shopping is simple because you need the same amount of all four. With contact ruled out, this arrangement has no touching pairs at all — four colours in equal numbers on this grid leaves plenty of room to keep them apart.

Dominant palette. 24 cream, 10 terracotta, 9 sage, 5 ochre. Cream carries the blanket and the other three punctuate it. Note that terracotta and sage were both asked for twenty per cent and came out at 10 and 9 — the odd square had to go somewhere. This is also the model where dye lots matter most, since half the blanket is one colour.

Exact inventory. 18, 12, 11 and 7 squares, because that is what was in the bag. It totals 48, which is what makes it usable; had it totalled 47 or 49, the grid would have had to change or another square would have been needed.

In all three cases the counts add up to 48 exactly. That is not a coincidence — it is the requirement. A layout where the squares do not fill the rectangle is not a layout.

Use the granny square colour layout generator

Everything above can be done on a floor with a notebook, and for a small blanket that is often the fastest way. For anything larger, or for a layout you want to be able to reproduce, the granny square colour layout generator does the placing.

It takes the grid, the palette and the frequency in whichever of the three forms above suits you — equal, weighted shares, or exact counts from your stash. It places the squares with same-colour contact ruled out if you want that, reports how many touching pairs are left rather than scoring the result, and tells you when what you have asked for is impossible rather than shuffling in hope. Every layout carries a seed, so the arrangement you liked is recoverable weeks later, and every square in the output is labelled with its colour's letter so the plan survives being printed in black and white.