top of page
  • Writer's pictureSally Khallash

How to find the optimum price for anything

You can't test your way to optimal pricing.

First of all it usually requires much more in terms of statistical significance than many people realise: if 2% of customers who take a look at your product ends up buying and you want to see if you can raise prices 10% then you'll need a whopping 78.039 customers through your test to achieve a statistically significant answer. Most companies don't see that kind of customer flow - especially not B2B - and even those that do will have a hard time keeping everything else constant to allow for a 'clean test'.

And we haven't even begun talking about price testing within your customer segments, which puts even higher requirements on your test size.

A related problem is what I call the 'Market Proxy'-Solution to testing price points: many companies (and their price consultants) simply see their whole market segment as a proxy for a large-scale live testing and try to gauge the price elasticity of their customer on how much business you and your competitors are getting relative to your price points. The biggest problem with the Market Proxy approach is that unless you are in a true commodity business like selling oil or electricity your customers will likely care who they are buying from and this preference isn't easily factored into the 'test'. Moreover: nobody in a group of competitors that are engaged in 'Market Proxy'-thinking has the incentive to try and raise prices too far away from the herd, which means that higher price points that are perfectly viable with your customers can remain undiscovered for a long time - or forever.

Second problem with testing is that your customers talk and a test large enough to give a meaningful result will have a near certain chance that customers will notice that you are floating different price points. Add a little social media - or some industry small talk - and you have a potential PR crisis on your hands.

Thirdly your customers will remember the prices you tested and be anchored by them - expecting to pay the same in the future.

Finally the whole idea of testing prices is run by the underlying rationale that you are somewhere on an 'optimisation hill', trying to find the top. Add a few percent to your price and see if your business improves. If it does: add a few percent more. If it doesn't: pull back.

This misses a key concept in marketing and business:

There might be a local optimum price in the market, which is usually where a group of companies or products are lumped together ('Market Proxy' above). But there might also be another positioning for your product or service in the marketplace that can allow you to multiply your price several times over. A new global optimum price.

Let's do a few examples to illustrate:


Starbucks entered the coffee market with a price point about 10x their competitors who were seeing coffee as an add-on side purchase. The competitors had, over time, competed the price of coffee down (coffee is a high margin business even on a bad day) and somewhere along the way their customers had forgotten that a cup of coffee can be both an indulgence and the center piece of a welcome break. Starbucks refocused on these things, added some deluxe packaging and reinvented the category.


When Jean-Claude Brouillet came to French Polynesia in 1975's he discovered black pearls and decided to introduce them to the American market. The problem was that nobody really liked black pearls - they were considered inferior to the traditional white Akoya pearls and demanded far lower prices.

But Brouillet was good friends with the New York jeweller Salvador Assael and convinced Assael to put the black pearls - which were relatively unknown in the US - on display next to his most exquisite and precious stones: diamonds, ruby's and sapphires. And he priced the black pearls accordingly - at twice the rate of Akoya pearls. And then they sold.

This positioning of black pearls amongst top jewellery items holds to this day.

Oil of Olay

Oil of Olay is owned by Proctor & Gamble - the 800 pound gorilla of fast moving consumer goods. But in the late 90's they were having problem with the cream: the customers who had initially loved the product when it was introduced in the late 1950's were now old women and, tellingly, people were starting to call Olay "Oil of Old Lady". Ouch. To make matters worse the cream was selling through pharmacies for $3.99.

Testing whether it would move for $4.49 wasn't really going to solve P&G's problem. They didn't need to revitalise the brand - they needed to reposition the product entirely.

They pulled Olay off of pharmacy shelves and focus grouped the following three price points with a whole new customer segment: women who shopped in department stores

  • $12.99

  • $15.99

  • $18.99

And the winner? $18.99 had the largest volume, followed by $12.99. Women who shop in department stores were used to paying somewhere between $25 and $400 for a cream. $18.99 hit the spot where it was considered cheap but still not so cheap as to signal it wasn't good.

P&G had invented what has since been known as the 'masstige' segment - the combination of mass-marget and prestige products. And almost 5x'ed their price point.


We have a case from one of our (happy but anonymous) customers: an European insurance company who was trying to market a pay check insurance product to its existing customers. For a few thousand euros a year, depending on your current salary, you could insure up to 90% of your current income in the case you lost your job.

The problem was that it didn't sell. When we interviewed some of their salespeople they complained that the product was too expensive and was taking far too long to explain to customers. They would usually pull it out at the end of a long conversation about home insurance, car insurance and similar items and try to add it onto the existing order. But the customers would almost never bite, so the commission paid salespeople were beginning to not introduce the product at all: a sure but certain death.

The first thing we did was notice that the Pay Check Insurance was the highest priced item in the basked of products the salespeople were bringing out - by a factor 3! Selling it would usually at least double peoples insurance budget.

So maybe the product wasn't expensive in itself - it was just being perceived that way because of what it was presented up against. The customers were anchored to the lower priced items and couldn't be pulled away.

So we took a look around and saw that our customer also owned a controlling stake in a mortgage lending operation - a much more promising situation. When people are buying a home they are already in a mindset where they are forking out €100.000's plus thousands in stamp duty and other fees.

Compared to this basket of purchases a few thousand euros a year seemed like a drop in a bucket. Also mortgage sales people are used to selling excruciatingly complex products to consumers. And the pay check insurance was a welcome addition as it fit in perfectly in the mix: if you are afraid of committing to years of high mortgage payments, why not insure your income stream and take away the risk?

Changing the positioning to speak to a different need of the same customers and shift the sales channel of the product made the original problems of 'too complex' and 'too expensive' vanish.


So far so good: I hope the above examples have convinced you that you can't test your way to the optimal price for your product. Price optimisation is fine - and absolutely necessary - when you are trying to optimise within your existing category and positioning in your market. But there are far better ways to do that than to simply throw out price A and B and then see what sells. More on that on another occasion.

But since category and positioning determines a disproportionally large part of your pricing power - often +80% for consumer goods - this is and should always be front-and-center in your thinking about price.

And here is how to do it:


Whether we are trying to establish a price point for a new product or service or trying to fix a product that isn't performing on either volume or margin our approach is the same: first make sure that we have the positioning right.

And this isn't just pie-in-the sky thinking where we try and figure out if we can charge 10 times as much for coffee if we re-invent the marketplace. Often our clients are selling their Black Pearls at Akoya prices - and their customers are unconsciously freaked out about it, wondering what's wrong with these black pearls since they are so cheap? Or simply buying happily. And silently. On other occasions we see an Akoya-quality product marketed at Black Pearl prices and have to engage our client in some soul searching.

Staying in the language of the image above, we want to make sure we are on the right 'hill' before we start to optimise for price.

To do that we go through these steps:


Market Fitting

The first thing we do is an iterative and somewhat circular process of what we call 'Market Fitting'. We try to find a way to position the product (or service) in a way so that there is a demand for it. Preferably a lot of demand.

The first step is to figure out who your customers might be, given what you're trying to sell. Use all the metrics that are relevant for your business - if you are in a B2C market, that might be Gender, Age, Income Level and so on. If you are B2B it might be Company Size, Location, Existing Invoicing Software and so on.

Then quickly move on: finding data about your potential customers - and correlations in that data - is a time consuming and entertaining task that has quickly diminishing returns. We usually find that the most important 5 metrics, as determined by you, carry +90% of the insights anyways. Don't spend your time on the next 500 just because the data is available to you.

Instead refocus your thinking not on WHAT your customers are, but on WHY they buy. At the end of the day a purchase is made not because a customer is a 36 year old mother of two with a €150.000 annual household income who has liked seven different Facebook groups about cats - she buys cat toys because she's worried that Mr. Smithers feels lonely.

The Harvard Business School professor Clayton Christensen has popularised the concept of 'Jobs to be done' - that a product fixes a specific problem a customer has. In his terms that we 'hire' a product to do a particular job for us. And that we as businesses should focus on the intentions of our customers in order to understand what jobs they need done and what business we are in.

What job did you hire that Condo to do?

When one of Clayton Christensens friends, Bob Moesta, used the 'Jobs-to-be-done' approach to find out why a developer couldn't sell new condominiums in Detroit to his target segment of people who were moving from larger to smaller homes, he found out that the customers were focusing on space for their old dining room tables - and the condominium kitchens were usually much too small for those. The customers were downsizing their lives but were focused on preserving the social gatherings and family time that they had had in their old homes. And this was hard to visualise without the dining room table as the stable center piece.

So Moesta figured out that the condominium developer was not in the business of providing affordable and luxurious condominiums to people who needed to downsize, but were instead in the business of 'Moving lives'. This was the job that people wanted done - and until the developer could fit the condominiums to contain as much of their old lives as possible they wouldn't sell. Check out the outstanding full article here on HBR.

The easiest way to do this is to go out and talk to as many customers as you can. Interview them with real curiosity about their lives (or businesses if you are in a B2B market).

In our framework we try to cover these four bases:

  1. Dreams & Aspirations: what does the customer really, truly want (and what would they pay big bucks to get).

  2. Fears & Pains: what is the customer afraid of. What is the biggest risk they just can't afford to run.

  3. Obstacles: what's in their way and why aren't they taking action?

  4. Situations: Focus on the specific, concrete situation that contains the job that needs to be done.

Covering these four topics with 10-20 customers usually gives you a lot of insight into what's going on. And yes: the insights are messy and qualitative and can't easily be fed into an excel model. That's okay, for now. We can quantify the Identities of our customers later when we're trying to optimise on our pricing hill - not when we're trying to find the right hill in the first place.

In the insurance example above the 'Situation' that contained the 'Job' was the home purchase. The customer was buying a home, taking on a great deal of risk and financial responsibility and was worrying about mortgage payments. The 'Job' was to fix the fear of missing those payments, so selling the pay check insurance right at the height of that fear - when the customer was about to commit - was the ideal Situation for that product.


The category of a product is simply to determine 'what it is'. If you invent a 0-kcal ice cream is it then a 'Dessert' or a 'Meal Replacement' or a 'Snack'? The choice you make might seem a little arbitrary until you realise that your choice of category really is a choice of your competitors. Would you rather compete with Ben&Jerry's, prescription NUPO shakes or Snickers chocolate bars?

Sometimes your category is very obvious, especially if you already have a portfolio of products in a given category and are simply expanding upon that. But we suggest giving it some thought anyways, just like Rolls Royce:

"Our competition is not necessarily vehicles, I can say airplanes, yachts, real estate, art, that’s where we are competing with. Our owners are putting down $20 million on a yacht. They may want to buy one or two Rolls-Royces to go along with it"

- Gerry Spahn, the head of communications for Rolls-Royce in North America

When you are shopping for extreme prestige your alternatives are not Rolls Royce, Bentley or Mercedes - because arriving in any of the latter two will have your billionaire friends asking why you are driving your wife's car today? And Rolls Royce would be missing out on a huge pricing opportunity if they were trying to compete with other cars. And they aren't: a Rolls Royce usually starts at 4x the price of a Bentley.

Your biggest asset in trying to determine your category is your research on your customers Dreams & Aspirations and their Fears & Pains. Ask yourself the question: why does the customer want this job solved, deep down? And then ask: what else in the world is currently being hired to do this job?

We usually call these deep reasons 'Benefits' and we run sort of a checklist to figure out what spots a product might hit:

So where Bentley's primary benefits are the Pleasure of driving and Self Actualisation through their "Be Extraordinary"-slogan, Rolls Royce is all about Status. In fact you don't drive a Rolls Royce - you are driven in one by your chauffeur.

Similarly the condominiums sold by Bob Moesta were being driven more by peoples fear of loosing Community and their need for Self Actualisation during this transition in their lives and less by their need for the Pleasure of a granite table top or larger bedrooms.

You might want to use a completely different framework to choose your category and that is fine. The important thing is to actively choose who to compete against. Because that choice will determine how and when you try and sell your product and who you are trying to position yourself against (coming up next).


If Category is determining who your competitors are then positioning is determining how you are going to compete. It is the full package of how you are promising your customer to get their job done - and not just whether you are priced high or low in your category.

Again your ressource is the customer research you did: mine your insights on both aspirations and fears as well as annoying obstacles and frustrations that your customer has.

And then choose one to focus on. Not two. Not three. Just one. Why? Because in positioning clarity wins. If you are renting out industrial machinery it's nice for your customers that you are both timely, trustworthy, focused on their needs and seventeen other things - but so is everyone else. By being everything you are being utterly unmemorable and none of your features are presented to your customer in a way that enforces any kind of real pricing power.

It's when you focus on one thing, like always being on time, that suddenly the customer starts to associate you with something specific and suddenly that element gains some pricing power. You are that company. This doesn't mean you should skip all the other features of course - it just means that it shouldn't be part of your core pitch.

This is hard: a lot of time and effort has usually gone into crafting products or services that cater to every little whim and fancy of your customers. And that is both good and necessary. But it isn't your position in your market.

See this one primary feature as the headline of your job application to your customer - if it's "Jane: Trustworthy, timely, hardworking and experienced" they are likely not going to bother reading your resume. But if you simply write "Jane - the hardest worker of them all" you are now much more prominently featured in the mind of your customer.


Once you have chosen a Category and a Positioning for your product or service it is time to validate it with your customers. And don't skip this: you might already feel that you've done a lot of work with your customers during your research phase, but that doesn't mean that your execution in terms of signalling to your customers that you've understood them is perfect yet.

Take the time to test. It pays of.

We usually recommend a qualitative testing in an environment as close to the real world as possible - if actual live testing in a commercial setting isn't possible. This might involve building a prototype product and getting it on the shelves in your actual retail channels. Or creating your service offering and asking a few trusted customers to be part of a pilot project.

While it is important that price is part of your offering here (never do free pilot projects), the price point isn't the primary research object here - you are not trying to find the top of the hill, you are trying to figure out if your choice of category and positioning has landed you on a hill in the first place.

Think of Oil of Olay: they tested three different price points but that was secondary to figuring out their new positioning in the department store market. And they actually ended up raising prices from $18.99 later on to as high as $50 for some product versions. Just to say that the important thing for P&G was to 'get on another hill' and not to try and hit the top of that hill in the first go. Because once they were they were confident that they could always optimise for price on that hill later on.

One of the reasons testing in a live environment is so important is because it brings you face to face with your competition in the context of your relevant sales channel and shows you exactly how your customers react. This is invaluable. And if you've made a mistake you would much rather make it in testing where it's easily fixable than in a full scale commercial launch.

And remember: this is more a qualitative testing than it is quantitatively. You are testing for demand, seeing if actual real life customers are crossing the chasm and making a purchasing decision. If they are, and you are satisfied with the results, then you're on a price hill and can start climbing once you launch.

What prices to test?

Work backwards from the sales channel you are presenting in: if you are positioning your 0-kcal ice cream as an ice cream in supermarkets next to Ben&Jerry's and 4 other ice cream brands then this is the category price range you are dealing with. Depending on your category you should stay within +/- 25% of that range - so if the ice creams cost from €3.00 to €7.95 then your possible prices should as a very general rule of thumb not go beyond €2.25 (discount positioning) to €9.95 (high end positioning).

You will obviously have an idea of what end of this range you are in and should test out 3-4 price points here spaced sufficiently apart and in-between competitor prices - i.e. test slightly above Ben&Jerry's, slightly below but still higher than the 2nd most expensive ice cream and then again between the 2nd and 3rd most expensive.

And that's it: choose the price final price that gives the highest profit to you and otherwise fits your overall strategy.


And THEN you can start optimising and do various testing to get to the top of your price hill - not before. How to do that without running into problems of underpowered statistics, price anchoring and customers that talk to each other - that's all for another post.

For now: I hope you enjoyed the read and I've managed to get you to think about pricing. Let me know in the comments what stood out the most to you or get in touch with me on

Ulrik Lehrskov-Schmidt

Partner, Behavioural Strategy (& former serial entrepreneur within education and FMCG)

BIO: Ulrik is a partner in Behavioural Strategy and holds a graduate degree in analytical philosophy from the University of Aarhus and has also studied an MA in Finance at Harvard University. Ulrik has worked with behavioural economics with a particular focus on financial decision making for a number of years.

He teaches master classes on commercial negotiations at Copenhagen Business School's branch of Executive Education and writes extensively on psychology, investing and financial decision processes - most recently in the upcoming book "Price Psychology" co-authored with Sally Khallash.

Ulrik has previously founded and sold four companies within real estate, FMCG and education.


bottom of page