Reverse-Engineering Marie's Blue Cheese Salad Dressing

Ever since I moved to Switzerland in 1991, after having lived almost twenty years in California, people occasionally ask, “Is there something you miss?” Well, I certainly don't miss the traffic jams on the grotesquely overloaded freeways, nor the weird people who made their home in the Sausalito post office and hit me up for spare change when I checked my mailbox at three in the morning, nor the moonbat politics and self-righteous politicians, nor the tax Nazis who make the IRS look like a huggable teddy bear. But, damn, if I don't miss the salad dressing. One of my first esculent enlightenments after moving to California was Marie's® Blue Cheese salad dressing, which is sold in the refrigerated section of the supermarket along with salad greens. (This was in the early 1970s; I'd never encountered Marie's before moving to California, but it is now available everywhere in the United States. I would usually include a link to the manufacturer's Web site, which is www.mariesdressings.com, but it is a horrific mess of Flash animations and ActiveX controls which I cannot in good conscience inflict on an innocent clicker of a link; if you want to visit it anyway, I've told you where to go.) I didn't even particularly like blue cheese dressing before stumbling onto this stuff—the various chemically-simulated horrors that come out of bottles with long necks which keep without refrigeration for centuries are as nothing compared to the bite of genuine chunks of blue cheese and the tart flavour of this stuff, which rapidly became a staple of my shopping list and one of the most difficult items to throttle back on when it came time to lose weight.

You can, if you wish, skip directly to the current recommended recipe.

The Real Stuff

Makers of mass-market food are understandably cagey about disclosing their recipes, but regulatory agencies do require them to list the principal ingredients in order of fraction of the product, and all significant minor ingredients, but there is no requirement to specify quantities which might be useful in developing an equivalent recipe. Still, we work with what we're given, so here's the information from the Marie's jar.

Now, the first thing that's obvious from this list of ingredients is what's immediately evident from the tiniest taste of the concoction: it's mostly mayonnaise. This is somewhat obscured by breaking out soybean oil, eggs, and egg yolks as separate ingredients, but if you put them together, you end up with ersatz mayonnaise (the genuine article being made with olive oil rather than the cheaper substitute of soybean oil). In fact, most recipes for mayonnaise include salt and vinegar (or lemon juice), and many also include mustard, which makes the flavour a bit more tart and contributes to the formation of the oil/water emulsion which gives the mayonnaise its creamy consistency. For example, here is the list of ingredients for the benchmark U.S. prepared mayonnaise, Hellmann's (sold under the Best Foods brand in the western U.S.).

Blithely ignoring, for the moment, precise quantities and proportions, we can simply “subtract out” the mayonnaise ingredients from those in the salad dressing, resulting in the simplified list of ingredients at the left. I've dropped out the xanthan gum, which is a flavourless natural additive used solely to increase viscosity; we'll wait and see if the result is too runny before worrying about ways to thicken it up.

Since, in practice, there's no reason to assume actual proportions of ingredients in the mayonnaise such as vinegar and salt are the same as in the salad dressing, we'll include them in the list when developing the recipe to permit adding additional quantities if necessary. A glance at the simplified ingredients reveals the secret which distinguishes Marie's from most other blue cheese dressings: the buttermilk and sour cream give an “edge” to the taste which perfectly complements the salty scrumptiousness of the blue cheese. Earlier attempts at replication without these essential components resulted in bland concoctions which tasted—not surprisingly—like blue cheese mashed up in mayonnaise!

Recipe Development

Finding a recipe which approximates the taste of the original was complicated by the fact that I'm working entirely from memory; I don't have access to the bottled glop to do the A/B comparison which is essential to arriving at a perfect match. On the other hand, I only need to satisfy my memory, not a panel of picky gourmets! I'm hoping that the publication of this document will motivate one or more enterprising cooks more talented than I who can compare against the commercial product to tweak the recipe toward a more faithful replica.

I also decided that, regardless of whatever compromise with fidelity it entailed, I was interested only in recipes which used commercial prepared mayonnaise as an ingredient. I am way too lazy to whip up my own from raw eggs and oil, and even something approaching the Platonic form of blue cheese salad dressing does not merit running the risk of a bout of Salmonella enteritidis. Besides, when it comes to fine-tuning the recipe, there are numerous varieties of mayonnaise on the market with different composition, and experiment may show one of them to be best.

Attempt 1

The ingredients for the first attempt are listed at the right. Since Roquefort cheese is sold by Swiss supermarkets in pre-packaged 100 gram slices, I decided to scale all the ingredients to one such slice. This results in a modest quantity of salad dressing which, if it ends up awful, doesn't take too many meals to use up before the next attempt. Quantities of ingredients you're likely to measure with a spoon are given both in units of metric teaspoons (5 ml) and tablespoons (15 ml) as well as volume in millilitres. Marie's is, of course, made with “Blue Cheese”, not Roquefort, which means it uses a knock-off product probably of domestic U.S. manufacture, not the genuine French AOC cheese made from the milk of sheep. Living in Europe, not far from the French border, I wouldn't consider using anything but the real thing, if for no other reason than it's not only genuine but also the most readily available. There are, of course, many other varieties of blue cheese produced both in France and around the world; it would be an interesting experiment to prepare and taste test salad dressing prepared from a selection of different kinds.

In this initial try, I decided to rely upon the ingredients in the mayonnaise for most of the minor flavours. The mayonnaise I was using in this and most subsequent experiments includes mustard and salt. It also includes vinegar, a detail I neglected to note, adding a substantial amount of vinegar to the mix myself.

Preparation couldn't be simpler. I just throw all the ingredients into a glass casserole dish and then chop up the Roquefort cheese with a metal serving spoon until it's in chunks, then stir everything together. You'll need to scrape the cheese off the sides and bottom of the dish where it tends to accumulate and blend it into the mix. You may also have to scrape off cheese which has stubbornly stuck to your spoon; a knife makes quick work of this. It is traditional to mix sauces like this with a whisk, but you don't want to do that when dealing with gooey blue cheese and already-thickened mayonnaise and sour cream; you'll end up with a blue cheese baseball bat which is a tedious mess to clean. You can use an electric mixer, food processor, or blender to mix the ingredients, but it's hard to achieve the right degree of chunkiness in the cheese. If you prefer a completely-mixed creamy salad dressing, however, that's the easy way to go.

So how was it? Execrable—the buttermilk and vinegar overwhelmed everything else and the result was a runny, sour mess with forlorn bits of blue cheese swimming in milky fluid. Well, after all, engineering is an art we learn from failure, and reverse engineering is no exception: onward.

Attempt 2

For the second attempt, it was clear I needed a lot more mayonnaise and/or sour cream to thicken up the consistency, so I doubled the amount of sour cream and almost doubled the amount of mayonnaise, keeping everything else the same. This was a big improvement; the result was still a bit thinner than I remember Marie's being, but not at all runny. The blue cheese was well dispersed throughout the dressing, so even sauce without chunks of cheese were well flavoured by the Roquefort.

I graded this attempt “getting there, but not right”, and decided to increase the creamy components and jazz up the spices the next time out. The result still seeming a bit too sour, and my having finally twigged to the fact that the mayonnaise I was using contained vinegar, I decided to halve the amount of vinegar.

Attempt 3

This batch was made during a trip to the British Isles, where Roquefort is sold in larger slices, which required adjustments to the other ingredients to compensate. Buttermilk wasn't available, so I did without. The British mayonnaise did not contain mustard, so I added 1/8 teaspoon (you aren't really going to measure 0.625 ml with a pipette, are you?) of dry mustard powder and, this time, some garlic powder. Although fresh garlic and prepared purée of garlic in a tube were available, I decided the powder would disperse the garlic flavour more evenly throughout the sauce.

Although the buttermilk was missed, the result had a delightfully unctuous consistency and tasted remarkably close to my memory of the real thing, which was quite surprising considering that the second ingredient on the label was missing! I decided the addition of mustard and garlic was a big improvement and decided to retain them in the recipe. This batch used a Waitrose “crème fraîche” produced in France (and which, despite the name, is in fact sour cream) which was extremely thick—much more so than the half-cream sour cream I used in attempts 1 and 2. I decided to use this kind of product in subsequent experiments.

Attempt 4

Based on my satisfaction with attempt 3, this time I added the buttermilk back in, but only 60 ml instead of the 100 I used before, in the hope of a less runny result, and I used the thick full-cream crème fraîche. I scaled back the major ingredients to their original proportions for the 100 g of Roquefort, but left the minor seasonings at the same levels I used for the 66% larger attempt 3.

This batch came out just great in my estimation; it may not be identical to the real thing, but it's close enough for me. After chilling an hour in the refrigerator, it was nice and thick, ideal as a dip for crudités as well as salad dressing. Even if you've well-mixed all of the ingredients initially, it's a good idea to stir the bowl once again after it's been sitting in the frigo for a while; the garlic powder and dry mustard don't immediately release their flavour when you add them to the mostly oily mix, and a second stir after a wait distributes their tang through the mixture.

Attempt 5

In the next attempt, I made one slight change. For the moment I'm satisfied with this recipe and have made it several times since without modifications. Note that the ingredients listed for Marie's include both “mustard flour”, the ground-up insides of mustard seeds, and “mustard bran”, the hulls from those seeds. Mustard bran has little taste of its own, but is a powerful thickening agent, and is presumably used primarily for this purpose. I originally considered it to be a specialised ingredient used in industrial packaged food production, unlikely to be found in supermarkets or in small quantities suitable for home cooking. I was delighted to find, however, that one can buy whole-grain ground mustard which includes both the flour and bran. In this attempt, I substituted that for the powdered mustard I had been using before and doubled the quantity to account for the presence of the mostly flavourless bran.

Comparison Tasting

In May 2006 I made a trip to the United States, and took advantage of the opportunity to bring back two bottles of the Real Stuff: one of the original “Chunky” recipe and one of the “Super”, which contains (according to the package) 25% more blue cheese. I made up a batch of the Attempt 5 recipe, let it age a couple of days to allow the flavours to meld, and compared them, initially by tasting small quantities directly from the bottle (or dish), and then on iceberg lettuce.

Initial evaluation? First of all, I have new respect for the experts who do comparison tasting of recipes. The power of suggestion is great in the human brain, especially in the deeper, more visceral parts associated with taste and smell. Just as one can persuade oneself that an audio clip is almost anything at all, the perception of flavour is strongly influenced by expectations, and the difficulty of comparison is magnified further in “chunky” recipes where each taste differs from the next. But still, one must judge, so here's a tip of the tongue report. In retrospect, I had underestimated how “mayonnaise-ey” the original Marie's recipe is. My approximation is much closer to the “Super” recipe, which I had never tasted before this comparison, than to the original which I gobbled regularly so many years ago. The difference between my recipe and the Marie's Super (to the limited extent of my discernment), is that the French Roquefort cheese I used is more strongly flavoured and more salty than the provenance-unspecified blue cheese used in the Marie's product. The difference is subtle: it's much more obvious when tasting the dressing directly than when served on a salad, and the distinction is more evident when the bit you're tasting contains a chunk of the undiluted blue cheese, not just the sauce. The whole-grain mustard I used appears to be ground more coarsely than the mustard flour and bran in Marie's, which gives the concoction a slightly speckled appearance kind of like Breyers vanilla ice cream, if you remember that (good grief—it appears to still exist; who'd have guessed?), and a very slightly “crunchy” texture. I don't find this at all objectionable, but if you do, substituting finely ground mustard powder should do the trick.

Having the three to compare side by side, I consider the “Super” Marie's recipe better than the original “Chunky” I tried to reproduce, and the difference between that and my recipe relatively small in a direct comparison and almost negligible when served on a real salad or as a dip for crudités. Still, I am an engineer, and attention to detail summons me back to the laboratory or, in this case, the kitchen.

The first obvious thing in the taste testing is that the blue cheese used in Marie's is substantially less strongly flavoured and less salty than the Roquefort I used, which can result in ambiguous results in taste testing of chunky recipes until you discover that the integrated taste of a spoonful depends on how many chunks of the straight cheese made it into your mouth. This is something which one can address only by changing the blue cheese in the recipe. I have laid in a supply of different kinds of blue cheese readily available in Central Europe, and I will taste them and try different recipes based upon them according to my evaluation of their similarity to the blue cheese used in the Marie's recipe.

I should observe here that Marie's have recently changed their formulation to jump on the latest U.S. food fad bandwagon and now shouts “0g trans fat!” in a yellow balloon on the label (compare the jars in the picture above with the one at the top of this document—who would have thought the absence of something could be so exciting?). I didn't pay any attention to this initially, but a few days after I did my initial comparison tasting, a reader wrote to thank me for posting the attempts to date, and remarked that with the change-over to reduce trans fat, “…the salad dressing is very disappointing. The new formula does not stay blended and the taste just isn't the same.” I found this very interesting, since I had noted myself when trying both the Marie's Chunky and Super I'd brought back on salads, neither seemed to be as good as I'd recalled Marie's to have been, but I'd initially put this down to faulty memory and aging taste buds (although the last time I'd had the real thing was less than a year before, and on that occasion it was as I'd remembered). If, indeed, the modified Marie's is inferior in flavour, it can't be used as a benchmark against recipes trying to reproduce the original, and since that product is no longer available, direct comparison is impossible. I suppose we should at least take comfort that Marie's didn't go and add peach fuzz to the goop.

Attempt 6

A few days after the initial taste testing, I was reading up on blue cheeses in Steven Jenkins's Cheese Primer, where he observes (p. 155), “I have never cottoned to the practice of blending Roquefort into a dressing for salad. For this purpose, any blue cheese, such as Danish Blue, will do, and at one-third to one-fourth of the price. The deep, full, spicy round flavor of Roquefort is denigrated when used in this manner. It deserves solo billing alongside a salad; then, both tastes are elevated, rather than diminished.”

Well, if you live in Switzerland, one thing you should never do is denigrate cheese, even if it comes from across the border! So, I decided to try a different kind of blue cheese in the next batch of SubMarie's, and picked up a variety when next I visited the grocery store, including the recommended Danish Blue (“Bleu Danois”—this was sold by the cut at the cheese counter and no brand name was in evidence, but according to Jenkins it is mass produced and generally consistent in quality), St. Agur, and Bleu de Bresse. I tasted these (and a few others, which weren't even close), and decided the Danish Blue was the closest to the blue cheese used in Marie's, with the St. Agur in second place, but closer to Roquefort than the Danish Blue.

Based on taste testing, I decided that from now on, my goal would be to reproduce Marie's “Super” recipe instead of the original “Chunky” because, having the opportunity to make a direct comparison, there's no question that the Super, with 25% more blue cheese according to the label (yet 10 fewer calories and 2 g less fat per serving—blue cheese, sinful as it is may be, still finishes second on the express lane to the afterlife compared to mayonnaise!) is without the slightest doubt the better salad dressing. Not only are there fewer calories in the Super, since it's more strongly flavoured, you may end up using less of it to obtain the same blue cheese bite on your salad.

With the revised target in mind, I made a batch of what has become my standard recipe modified as indicated at the right. The only changes were substituting the Danish Blue cheese for Roquefort and using Colman's powdered mustard instead of the ground whole-grain mustard I used before. The latter was because I noticed the whole-grain resulted in dark specks and a slight crunchiness absent in the prototype. I should have reduced the quantity of ground mustard to 1/8 tsp to compensate for the absence of mustard bran, but I didn't and I doubt if anybody would notice the difference.

After leaving this batch to mellow overnight in the frigo, I compared it with the Marie's Super and concluded that it was almost indistinguishable in direct A/B taste testing. At the same temperature, my recipe was somewhat more thick than Marie's, but both had flecks of blue cheese distributed about equally through the bulk. Unlike my recipe made with Roquefort, which was distinctly more green than Marie's, this batch was just about the same shade of blue. Trying to be as fussy as possible, I noted the following slight differences:

• Marie's seems to have slightly more buttermilk bite than my recipe. I will increase the buttermilk in the next “build”. This will restore buttermilk as the second largest ingredient in the recipe, in accordance with the list on the Marie's package, and also reduce the viscosity, which I was previously trying to maximise after an initial bout of runniness, but in which I may have overshot the goal in the latest rounds.
• My recipe still seems slightly more salty. I will delete the small amount (1/8 tsp) of added salt the next time.
• I will reduce the quantity of ground mustard. I doubt anybody can tell the difference, but I suspect the buttermilk would benefit from less competition.

We are definitely into the domain of fine tuning here. The recipe above, with Danish Blue, makes a better salad dressing than the Roquefort in my estimation, but if you're one of those people who blesses every bite with the sodium ion dispenser, the difference may not be all that apparent. There is still a little “crunchiness” to this recipe compared to Marie's: this is due to inherent granularity in the Danish Blue cheese which is not present in that used in the Marie's dressing. The difference is apparent only if you taste the dressing directly (which can lead to drinking it with a straw, intravenous injection, and even more ignominious compulsions—step back from the brink, while there's still time), and imperceptible when you use it on a salad or as a dip.

I suspect that if you want to approximate the original “Chunky” recipe, you can make up a batch of my emulation of the “Super” and give it a good squirt of mayonnaise (and maybe some sour cream) to increase the volume by about 20%, and you'll end up with something close.

Attempt 7: Current Best Practice Recipe

Based on further taste testing and adjustments, I arrived at the recipe at the right, which is now my standard; I have made numerous batches of it and find no need for further adjustments. I have re-scaled the recipe around 150 g of “Castello” Danish Blue cheese because it is sold pre-packaged in that quantity in Swiss supermarkets. I find the 50% larger batch size more convenient; I like this stuff so much that I'm guaranteed to finish it before it goes all entropic.

In this iteration, I have gone heavier on the sour cream (continuing to use thick “crème fraîche” variety) and buttermilk at the expense of the mayonnaise; I find this, along with the substantially more vinegar in this iteration, increases the “bite”, which compensates for the milder taste of the Danish Blue cheese. I've deleted the added salt and kept the mustard and garlic powder the same even though the rest of the recipe has been scaled up; they are such minor components of the composite taste that you could probably double them and not notice the difference.

After making this recipe a number of times, I have made a small refinement in the way I combine the ingredients. First, I add the ground mustard and garlic powder to the mixing bowl, then the vinegar and buttermilk. I then stir until the mustard and garlic dissolve in the thin, acidic liquid. Next, I add the mayonnaise and sour cream, and then blend everything to a smooth consistency. Finally, I add the cheese and chop it up into a mix of finely dispersed bits and larger chunks and finally mix well. I find that two serving spoons do a fine job of chopping and mixing the cheese, the second used for scraping the first when a glob of cheese accumulates on it. With so much buttermilk in the mix, the result will be somewhat runny when you're done stirring it, but if you leave it in the refrigerator overnight, it will thicken up nicely. Be sure to stir it again before serving.

The Bottom Line

Like all too many yummy things, this is just about the polar opposite of a low-fat, low-calorie health food. One batch of this stuff (based on the 150 g Danish Blue cheese recipe above) adds up to almost 1600 calories—comparable to the entire daily calorie requirement of a smallish adult, and almost all of it is from fat. Each tablespoon you glop on your salad rings up about 80 calories on the Eat Watch, so bear in mind that a little goes a long way, and that it's all too easy to transform your healthy garden-fresh salad into an artery-clogging calorie bomb through the liberal application of this creamy concoction.

It is possible to substantially reduce the calories in this recipe by substituting low-fat mayonnaise and sour half-cream for their fat-city congeners. Doing so (using products from the same manufacturers for comparison, although they're all about the same), reduces the calories from the sour cream to 171 and the mayonnaise to 267, for an overall reduction to 1127, or about 30% less, which reduces the calories per tablespoon to around 57. It is interesting to note that Marie's make their own “Lite” version, with a somewhat intimidating list of ingredients and a greater (44%) reduction in calories. Double-blind taste testing of the regular and reduced calorie versions of both the Marie's products and the SubMarie's approximations presented here would be an interesting experiment for a cuisinerd dinner party.

A Matter of Taste

While Attempt 7 is the recipe I have settled on and use routinely, in my estimation any of the recipes from Attempts 5 through 7 produce an excellent salad dressing, and the choice among them is a matter of personal preference. If you prefer a more pungent and saltier concoction, by all means try the Rouquefort-based Attempt 5. Attempts 6 and 7 both yield a milder flavour (I believe the blue cheese in Marie's is more strongly flavoured than Danish Blue, but closer to it than to Roquefort), with Attempt 6 leaning more in the direction of mayonnaise and Attempt 7 toward buttermilk and sour cream. I find that the latter compensate for the milder blue cheese, but I'd encourage you to try them all, settle on the one you prefer, and feel free to experiment further!

Keep in mind that there is substantial difference in the flavour of various brands of commercially prepared mayonnaise, sour cream, and buttermilk, and that the formulation of these products is adjusted according to the regional preferences of where they are sold. If you find, for example, that the Attempt 6 recipe is too heavy on the mayonnaise and Attempt 7 is overwhelmed by the sour cream, try varying the proportion between them until you achieve the correct balance. (The “crème fraîche” that I buy in Switzerland is very mild, nothing as sharp as sour cream I recall from the U.S. If your sour cream is really sour, you probably need to reduce the proportion. Similarly, buttermilk varies substantially both in the strength of its flavour and consistency; if you end up with a runny mix that's dominated by the flavour of buttermilk, you'll want to reduce the amount you use.)

Back to FourmiFood

Fourmilab Home Page

by John Walker February, 2006 Revised September, 2007

This document is in the public domain.