Healthier Glazed Donuts (24 → 43): Baking Them Was the Smallest Part

A glazed donut is one of the few foods nobody pretends is health food. That is part of the appeal. You are not being sold anything - it is fried enriched dough in a sugar shell, it is warm, it is gone in four bites, and everyone at the table knows exactly what they signed up for.

So we built the standard version in Eat Well Planner - a classic yeast dough, deep fried, dipped in a proper thick glaze - and ran it through the app's Nutrition Score. It came back at 24 out of 100. Then we rebuilt it one verified swap at a time and got it to 43.

That is a real 19-point move, and it is also, we should say up front, the honest ceiling. A donut that scores 80 is not a donut. What surprised us was where the points came from. The swap everyone recommends first - bake them instead of frying them - was worth zero points when we tested it on its own. The two changes that actually moved the number were the flour and the glaze, and neither one is the change people expect.

The original: what's in it and how it scores

We scored a standard homemade yeast donut, the kind that shows up in every "just like the donut shop" recipe. The dough: 3 1/2 cups all-purpose flour, 1 cup whole milk, a packet of active dry yeast, 1/4 cup granulated sugar, 2 eggs, 4 tablespoons butter, 1 teaspoon salt. Fried in vegetable oil, of which each donut absorbs a little over 8 grams. Then dipped in a real glaze: 2 cups powdered sugar thinned with 3 tablespoons milk and a teaspoon of vanilla. Twelve donuts. Every number below is per donut.

Nutrition Score: 24/100. Per donut: about 369 calories, 6.1 g protein, 1.1 g fiber, 14.4 g total fat, 4.1 g saturated fat, 240 mg sodium, and 24.1 g of added sugar, which the app flags as very high at 91% of a day's reference. For context, the FDA sets the Daily Value for added sugars at 50 g, based on the Dietary Guidelines advice to keep added sugars under 10% of daily calories. One donut is roughly half of that.

Here is the part that reframed the whole post for us: the frying oil is not what is wrong with a donut. Vegetable oil is mostly unsaturated, so those 8 grams add about 75 calories and barely touch the saturated fat - the 4.1 g of saturated fat is almost entirely the butter in the dough. The oil was actually contributing something the score liked, which we will come back to.

What genuinely holds the donut at 24 is narrower and duller than "it's fried": a quarter cup of sugar in the dough plus twenty grams of powdered sugar on top, and 3 1/2 cups of flour that has had its fiber milled out of it. Fiber comes in at 1.1 g, against a Daily Value of 28 g. The recipe scores 1 plant point against the app's 30-a-week gut health target. Sodium is not flagged at all. Neither is saturated fat.

It is not all bad news, either. The donut is an excellent source of thiamine (36% DV), folate (38%), iron (28%) and riboflavin (20%), because American all-purpose flour is enriched by law. It is not a nutritional void. It is a food with one very specific problem.

The upgrades

We applied these in order and re-scored after each one, so every delta below is what that swap was worth at that point in the build.

1. Swap half the flour for white whole wheat

• Swap: 3 1/2 cups all-purpose flour → 1 3/4 cups all-purpose + 1 3/4 cups white whole wheat
• Score impact: +9 points (24 to 33) - by far the biggest move in the post
• What it changes: Fiber from 1.1 g to 3.0 g a donut, nearly triple, plus the magnesium and phosphorus that ride along with the bran
• What it costs you: Almost nothing at 50%. A slightly deeper color and a faintly nuttier flavor that the glaze covers anyway.

Nine points from one bag of flour. Refining wheat, as Harvard's Nutrition Source puts it, strips out the bran and germ and with them more than half of wheat's B vitamins, 90 percent of the vitamin E, and virtually all of the fiber - and the same page notes that women eating two to three servings of whole grains a day were about 30% less likely to have a heart attack or die of heart disease over a decade than those eating less than one serving a week. Putting half of it back is the highest-leverage thing you can do to a donut.

Use white whole wheat rather than standard red whole wheat. It is a true whole grain flour milled from a paler, milder wheat, so it carries the same bran without the assertive wheaty bitterness, and in a delicate enriched dough that difference matters. Two technique notes: whole wheat flour drinks more liquid, so if the dough feels stiff while kneading, add milk a tablespoon at a time; and give the dough its full rise, because bran physically interrupts gluten and a rushed whole wheat dough bakes dense. If you only have red whole wheat on hand, our all-purpose flour substitutes guide covers the ratios.

2. Cut the glaze back and thin it

• Swap: 2 cups powdered sugar with 3 tablespoons milk → 3/4 cup powdered sugar with 2 tablespoons milk
• Score impact: +3 points (33 to 36)
• What it changes: Added sugar from 24.1 g to 11.7 g a donut - out of the app's very high band and into high; 49 fewer calories
• What it costs you: The opaque white shell. This glaze sets thin and translucent, more lacquer than frosting.

Halving the biggest sugar source in the recipe is worth three points, and the gap between the size of that change and the size of the reward is worth sitting with. The score does not pay you much for a smaller sin; it pays you for adding something. Cutting the sugar in half moved the number a third as far as adding fiber did.

It is still the change we would make first if you only make one, because of what it does off the scoreboard: 12.4 fewer grams of added sugar per donut, which is most of a day's difference if donuts are a regular thing in your house. Whisk the glaze thinner than feels right and dip the tops only - a thin coat on a warm donut sets into a crackly shell rather than a sweet slab, and you taste more of the dough underneath. Do not be tempted to replace the powdered sugar with honey or maple syrup: the app counts those as added sugar in exactly the same way, and they are wetter, so the glaze never sets properly. Our guide to white sugar substitutes goes through which of those swaps are genuine nutritional upgrades and which are just a different sugar.

3. Bake them instead of frying them

• Swap: Deep fried → baked at 375°F for 8 to 10 minutes
• Score impact: +3 points (36 to 39) - and +0 points when we tested it on the unmodified original
• What it changes: Calories from 313 to 238; total fat from 14.5 g to 6.1 g; saturated fat from 4.1 g to 3.0 g
• What it costs you: The crust. This is the real trade in the post, and it is not a small one.

We scored this one twice because the result was so counterintuitive. Take the original recipe exactly as written and bake it instead of frying it, changing nothing else: the score goes from 24 to 24. Not one point. Calories drop from 369 to 294 and total fat drops from 14.4 g to 6.1 g, and the score does not move, because the fat that came out was unsaturated fat it was not penalizing in the first place. If anything the app liked the fried version's fat profile - the vegetable oil was supplying 41% of the Daily Value for omega-3 ALA, and baking took that to nothing. Later in the build, once the glaze sugar and the refined flour had been dealt with, the same swap was worth three points, mostly because 75 calories is a bigger proportion of a lighter donut.

So bake them - but bake them for the calories and the fat, not because frying is the villain. And be honest with yourself about what you get. A baked yeast donut has the soft, pillowy crumb of a fried one and none of the thin crisp shell that hot oil sets on the outside in ninety seconds. It lands closer to a very good enriched roll than to a donut shop donut. If you would rather keep frying and just want a better fat for it, our vegetable oil substitutes guide covers smoke points and which oils hold up.

4. Add two tablespoons of ground flaxseed to the dough

• Swap: Add 2 tablespoons (20 g) ground flaxseed to the dry ingredients
• Score impact: +4 points (39 to 43)
• What it changes: Omega-3 ALA from 0.05 g to 0.37 g a donut - back to an excellent source at 23% DV; fiber up to 3.4 g; a third plant point
• What it costs you: Nothing detectable. Twenty grams across twelve donuts is under two grams each.

This upgrade exists because of what the last one took away. The frying oil was the donut's omega-3 source, and once the fryer is gone something has to put it back. Ground flaxseed does it far more efficiently than oil ever did: the USDA notes that each tablespoon of ground flaxseed contains about 1.8 grams of omega-3 fatty acids, along with fiber and protein, in a quantity far too small to taste.

It has to be ground. The same USDA page is blunt about it: flaxseed eaten whole is more likely to pass through the body undigested, which means the body does not get all of the nutrients and bioactive components. Buy whole seeds, grind what you need, and keep the rest in the freezer, because the oil in pre-ground flax turns rancid quickly. Four points for a tablespoon of pantry seeds is the best return in this entire post.

What we kept, and where we stopped

The butter stays. All four tablespoons of it, and it is now the largest source of saturated fat in the donut. An enriched yeast dough is enriched by fat, and butter is what makes the crumb short and tender rather than bready - in a baked donut, which has already given up its fried crust, taking the butter out too leaves you with a sweet dinner roll. The eggs stay for the same reason. The salt stays, at 240 mg a donut, which the app does not flag.

And the glaze stays, which is the line we drew deliberately. Dropping it entirely takes the score from 43 to 52 - nine more points, more than any single upgrade above earned. We are not doing it, because a yeast ring with no glaze is not a glazed donut, and a makeover that deletes the word on the label has not improved the recipe, it has changed the subject. Forty-three is what a glazed donut can honestly score.

The healthier version: baked glazed donuts

Nutrition Score: 43/100. Makes 12 donuts.

Dough

• 1 cup (240 ml) whole milk, warmed to about 110°F
• 2 1/4 teaspoons (1 packet) active dry yeast
• 1/4 cup (50 g) granulated sugar
• 1 3/4 cups (220 g) white whole wheat flour
• 1 3/4 cups (220 g) all-purpose flour
• 2 tablespoons (20 g) ground flaxseed
• 1 teaspoon (6 g) fine salt
• 2 large eggs, room temperature
• 4 tablespoons (57 g) unsalted butter, softened

Glaze

• 3/4 cup (90 g) powdered sugar
• 2 tablespoons (30 ml) whole milk
• 1 teaspoon vanilla extract

1. Stir the yeast and 1 tablespoon of the granulated sugar into the warm milk and let it stand 5 to 10 minutes, until it foams.
2. Whisk both flours, the ground flaxseed, the remaining sugar and the salt together in a large bowl.
3. Add the yeast mixture and the eggs and mix to a shaggy dough. Work in the softened butter a tablespoon at a time, waiting until each addition disappears before adding the next.
4. Knead 8 to 10 minutes by hand, or about 6 minutes in a stand mixer with the dough hook, until the dough is smooth and just slightly tacky. Whole wheat flour absorbs more liquid than white, so add milk a tablespoon at a time if it feels stiff.
5. Cover and let rise in a warm place for 1 to 1 1/2 hours, until doubled. Give it the full time - a rushed whole wheat dough bakes dense.
6. Roll the dough 1/2 inch thick on a lightly floured surface. Cut with a 3-inch round cutter, then a 1-inch cutter for the centers. Re-roll the scraps once.
7. Set the donuts on parchment-lined sheets, cover loosely, and proof 30 to 45 minutes, until visibly puffy and slow to spring back when poked.
8. Bake at 375°F for 8 to 10 minutes, until the edges are just golden. Pull them the moment they color - overbaking is the only real failure mode here, and it makes them dry.
9. Whisk the powdered sugar, milk and vanilla until smooth. It should ribbon off the whisk rather than mound; add milk a teaspoon at a time if it is too thick.
10. Dip the tops of the still-warm donuts and set them on a rack for the glaze to set. Best within a few hours of baking.

Per donut, original vs. improved:

• Nutrition Score: 24 → 43
• Calories: 369 → 247
• Protein: 6.1 g → 6.1 g
• Fiber: 1.1 g → 3.4 g, a good source at 12% DV
• Added sugar: 24.1 g (flagged very high) → 11.7 g (flagged high)
• Total fat: 14.4 g → 6.8 g
• Saturated fat: 4.1 g → 3.0 g
• Omega-3 ALA: 41% DV → 23% DV, still an excellent source
• Plant points: 1 → 3

Swaps we tested and skipped

The air fryer. We scored the finished dough as an air fryer version with a light spray of oil, since that is the first thing anyone asks. It came in at 42 against the oven's 43 - one point worse, because the spray puts a little fat back without buying anything for it. Texturally it is a wash at best: circulating heat browns the outside faster than the inside sets, so you get a tougher skin and a drier crumb unless you watch them closely. Use the air fryer because the oven feels like overkill for six donuts, not because it is healthier.

Greek yogurt for some of the milk. A standard move in lightened baking, and here it did precisely nothing: 43 before, 43 after. It nudged protein from 6.1 g to 6.8 g and shaved a little fat, which is not enough to move the score or to justify the tang it puts into a plain glazed donut.

Applesauce for the butter. This is the one we rejected against the numbers, and we want to be clear that it was a judgment call. Replacing the butter with 1/4 cup of unsweetened applesauce took the score from 43 to 47 and cut saturated fat from 3.0 g to 0.9 g. Four points is real. We still would not do it, because the applesauce version is a bready, faintly gummy ring - the fat is doing structural work in an enriched dough that pectin and water cannot do. Make it that way if saturated fat matters to you medically, and go in knowing what you are trading. Our butter substitutes guide covers the swaps that hold up better in enriched doughs.

100% white whole wheat. Taking the flour all the way whole grain scores 46, three points above the half-and-half version, with fiber at 5.4 g a donut. It is a perfectly reasonable option and the one we would pick for ourselves. It is not the default here because the crumb goes noticeably denser and chewier, which is a real cost in a food whose entire point is being pillowy. Start at half and go up if you like where it lands.

Or just make them smaller: the honest verdict

Because we knew someone would ask, we scored the original fried recipe cut into 18 smaller donuts instead of 12. That donut comes in at 246 calories and a score of 25.

Put that next to the improved version at 247 calories and a score of 43, and the choice gets unusually clear. On calories they are the same donut. What separates them is everything a calorie count cannot see: 3.4 g of fiber against 0.7 g, 11.7 g of added sugar against 16.1 g, half whole grain flour against none. So the answer depends on how often you are doing this. If donuts are a twice-a-year thing, make the real ones, make them a little smaller, eat them hot out of the oil and enjoy them completely - the occasional fried donut is not the thing standing between anyone and good health. If donuts are a Saturday morning habit, the baked version is the better weekly donut by a wide margin, and it is still recognizably a donut.

Make it fit other diets

For dairy-free, unsweetened soy milk swaps in for the whole milk one for one and a good plant butter handles the 4 tablespoons; the dough and the glaze both survive it intact. For vegan, the eggs are the real obstacle rather than the dairy - they are doing structural work here, so that version is a rebuild rather than a substitution. Gluten-free is the one we would not attempt as a straight swap: a yeast donut is a gluten structure, and a 1:1 gluten-free blend gives you a cake donut at best.

Rather than guessing at ratios, drop the recipe into Eat Well Planner and ask Recipe Variations for the version you need. It rewrites the whole recipe with the swaps worked in, and you can score the result before you commit two hours of proofing to it.

Do this to your own recipes

Every number on this page came out of the app's Make It Healthier feature and its Recipe Playground. You import a recipe - from a URL, an Instagram reel, a YouTube video, or a photo of a handwritten card - and the app scores it, then proposes ingredient swaps with the nutritional gain verified by a calculator before the suggestion ever reaches you. In the playground you change one thing at a time and watch the number move.

That is the only reason we found out that baking a donut was worth zero points and a tablespoon of flaxseed was worth four. Nobody would have guessed that order. Try it on the recipe you would never give up - you will usually find the real fix is smaller, more specific, and less painful than the internet told you it would be.