I love these cookies from King Arthur Flour for a variety of reasons, including that they're delicious and they only need five ingredients, which I always have in my pantry. I particularly love the thumbprint version where you press a dent in the center and fill with jam before baking.

I do love the way they taste but I have friends who are vegan and I'd like to try substituting the butter with coconut oil. I've had really great success with cookie recipes that make use of coconut oil (like Brave Tart's rolled sugar cookies which use both butter and coconut oil) but this is a flour-based cookie and it's still using part butter.

So, is the recipe below likely to have issues if I sub the oil for butter? With only the almond flour, butter and sugar really doing much, my concern is that they won't bind together as well or that they may spread more. One of the comments from KAF said that the butter acts as the binder, so I'm not sure if the coconut oil binds as well... particularly since one of the comments says:

Loved this cookie and the low carb aspect! Tried it again, replacing butter with coconut oil. Flavor was great, but the cookies fell apart. Any advice on the ratio of fat I should have used.

So, anyone have any thoughts on how to avoid this? More/less coconut oil? A different butter substitute (I'm not a huge fan of margarines and their fake butter flavor)?

I'll probably try it out if no one has an answer. I don't have any refined coconut oil on hand (not interested in adding the coconut flavor), which is why I'm asking instead of just trying it. Fortunately, the batch size is small, so it's easy to do test batches.

Gluten-free Almond Flour Shortbread Cookies


  • 3 3/8 ounces almond flour
  • 1 1/2 ounces softened butter
  • 3/4 ounce confectioners' sugar
  • 1/8 teaspoon salt
  • 1/2 teaspoon vanilla extract


  1. Preheat the oven to 350°F. Line a baking sheet with parchment paper.
  2. Mix all of the ingredients in a small bowl until a cohesive dough forms.
  3. Scoop 1" balls of dough onto the prepared baking sheet; a teaspoon cookie scoop works well here. Arrange the balls of dough about 1 1/2" to 2" apart.
  4. Use a fork to flatten each cookie to about 1/4" thick, making a crosshatch design.
  5. Bake the cookies for 8 to 10 minutes, until they start to turn light golden brown on top.
  6. Remove the cookies from the oven and cool them on the pan for 10 minutes. Transfer them to a rack to cool completely before serving.
  • My experience with coconut oil was fairly one-to-one substitution worked, in brownies. Not sure about these cookies. Mom used to use margarine for all her baking, is there vegan margarine? Commented Apr 18, 2019 at 17:49
  • 2
    butter is fat + water ... so you might need to reduce the coconut oil slightly, and add in some water. Some brands of margarine actually have water in them so it's easier to use them as butter replacements
    – Joe
    Commented Apr 18, 2019 at 17:49
  • 2
    Ok, so I did a little experiment (still not answer-worthy because of more substitutions), using coconut flour and coconut oil. The dough was incredibly crumbly, like the coconut flour “soaked up” the oil. Added a tiny bit of water, with no significant difference. Didn’t dare to add more than a scant tablespoon. Still couldn’t shape flat cookies, because they would crumble immediately. Baked little balls. Still fall apart if you look at them sternly.
    – Stephie
    Commented Apr 18, 2019 at 18:10
  • 1
    You probably need longer chain fatty acids than coconut oil provides. Palm oil is a possibility: en.wikipedia.org/wiki/Palm_oil#Fatty_acids There have got to be others. Commented Apr 18, 2019 at 23:34

1 Answer 1



Pure coconut oil is not an adequate replacement for butter in this recipe. However, coconut oil and water appears to work as a substitution when using ratios between 4:1 and a 5:1 coconut oil to water by mass.

Cookies produced with pure coconut oil fall apart when touched. Ratios of 4:1 and 13:2 coconut oil:water worked in my test.

However, the substitutions tested did not produce an exact match for the recipe produced with butter. (h/t rumtscho for catching the omission of reasons to predict this outcome).

Empirical Tests

Test Procedure

I used a batch size that was 1/3 that of the included recipe, or

  • 32g Almond Flour
  • 7g Confectioner's Sugar
  • Equivalent* of 1/2 ounce butter
  • A pinch of salt (quantity too small to measure)

The butter equivalents tested were:

  1. 15g Coconut Oil
  2. 13g Coconut Oil and 2g water
  3. 12g Coconut Oil and 3g water
  4. 1 tbsp butter, as marked on the butter package (16g in this trial)

Batches were mixed in the order specified in the butter equivalents list on the same equipment. All batches were placed on a single tray and baked at the same time

Two of the batches were divided into 6 cookies instead of 5 to make them distinguishable from the other batches.


Coconut Oil, No Water

Cookies produced with a 1:1 replacement of butter with coconut oil were extremely fragile. I believe this reproduces the result reported in the complaint cited in the question.

I decided to give up trying to transfer them to a cooling rack after fracturing all cookies in the batch multiple times, so I do not have post-cooling observations for this batch.

Coconut Oil and Water, (13:2) ratio

Cookies hold their shape when handled but crumble readily when chewed. Compared to butter cookies, these are slightly harder along the bottom, and have more snap (it takes some effort to break by bending).

Coconut Oil and Water, (12:3) ratio

I was not able to notice a difference between cookies of this batch and cookies of the 13:2 coconut oil to water ratio batch.


Note that 16g of butter (instead of the planned 14g) was used, possibly skewing comparisons.

Cookies produced with butter produced results that were between that of the coconut cookies with and without added water; they crumble easily in the mouth but not during handling, and can be broken easily by hand.

Conjecture and References

Some wheat shortbreads are made with oil and water instead of butter, so having an oil and water emulsion as a replacement for butter may not be critical; therefore, it may suffice to simply replicate the fat:water ratio of butter, which is approximately 5:1 by mass.

The existence of vegan shortbread recipes that use vegan butter substitutes suggests that commonly-used vegan butter substitutes may also be acceptable; building on the earlier conjecture, it may suffice to replicate the fat:water ratio of a commercially-available vegan butter substitute. Earth Balance has a 4:1 fat:water ratio by mass.

The original conjecture before empirical tests was that coconut oil and water added in a 5:1 or 4:1 ratio by mass would be an acceptable substitute for butter.

Dairy Butter Characteristics

Dairy butter has a nonzero amount of water. This American source reports about 16% water and 82% fat for American butters.

Using these numbers, this recipe contains about 42.5g butter, which contains about 6.8g (or ml) water. Substituting butter with pure oil removes 6.8 out of 9.3ml of non-oil wet ingredients (2.5ml from the vanilla extract); if non-oil wet ingredients are important, this is a significant change.

As noted by rumtscho in a comment, butter is an emulsion of water and fat and may behave differently in a recipe compared to adding an equivalent amount of fat and water. A difference between butter and coconut oil with added water was observed in empirical tests.

Vegan Butter Substitutes

Earth Balance original spread (easier to read form here) has a significant amount of water. From the calorie counts (assuming 9 Calories per gram of fat) and the fact that ingredients other than oil and water are listed as <2% ingredients, Earth Balance appears to be around 20% water.

Other Shortbread Recipes

So far, I have not found a directly-comparable recipe.

I found three recipes for wheat shortbread; one without any added water, another that adds water in the form of syrup, and a third that suggests adding plant milk as a binder but discourages it. The first recipe is evidence against the inference that added water is strictly necessary. The third recipe is weak evidence that it is not necessary to have all fat and water emulsified beforehand.

However, almonds and wheat flour have considerably different characteristics, so it may not be straightforward to generalize from successful wheat flour recipes.

  • 2
    I am impressed with the effort you put in that answer. However, I am not comfortable with your assumption that one can just "add the water". The water in butter is emulsified, so adding water plus fat behaves very differently in dough than adding butter.
    – rumtscho
    Commented Apr 19, 2019 at 12:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.