XX.VI. Demand

The 6th system architecture pattern: achieving product-market fit
Dr. Albert Künstler

ow that we have discussed the common processes, let us consider the most popular system structures that you may encounter over the course of your quest.

Since we are talking about for-profit research, here is an obvious question: Why did the sponsors decide to spend money on this quest?

The answer is pretty clear: they want to increase their profits. Profit arises from the sale of products and services, which require a perfect understanding of effective demand. Therefore, the central topic in many research studies is understanding the structure of demand, as in what consumers need and what they are willing to pay for.

Research landscape of the Demand structure

To understand the demand you have five methods:

  1. Collect information on what customers buy from your competitors
  2. Hire an expert with deep market knowledge or find secondary sources of aggregated market analytics
  3. Ask customers about their needs
  4. Offer them something and see if they are interested
  5. Sell them the actual product and see how they pay and use it

The last method gives you most knowledge about demand, because you have full control over your product and complete information on real customer behavior, but it requires a mature product and established customer acquisition process to make statistically significant conclusions. Thus, you might need to use a full spectrum of the above methods.

If you want to do so and have the full picture of the demand landscape, here are the main research spaces: Customer needs, Customer segments, Channels, Competitors, Product, Messages, and Profit.

Customer needs

Customer needs is the list of all possible desires and requirements that satisfy the customer demand. This is an exhaustive list that describes not just one customer, but customers in general. No individual customer will expect all these needs to be satisfied, but some needs will be more statistically important than others. This prioritized list is extremely valuable and must be maintained with great diligence.

In our future chapters we will introduce the concept of customer genotypebased on this list of needs, that will be our fundamental tool for strong problem solving.

Customer segments

Not all customers are the same. It makes sense to differentiate them into several groups based on their behavior and needs. Customer segments, frequently referred to as personas, will require different approaches in both selling and applying the product. The list of needs is the same for all segments, but the priorities differ. Some needs may be central for one segment but insignificant for another.


Customers are not evenly distributed. Some locations, or channels are more densely populated with customers, so these channels may bring more buyers. That is also true for segments. Some channels have higher representation of a certain segment over the others.


Competitors are companies in the same industry that sell similar products to your potential customers. Competitors are a wonderful source of insights about demand, as they already did their research, already designed the products that satisfy the customer needs, and already identified the channels with the relevant audiences. However, they are your rivals and they are not very happy with your presence in their market. And of course, they will try to protect their trade secrets.


This is the offering of your quest giver. In an ideal world, the product must satisfy the demand that you are now trying to explore. That being said, the variations are countless, and the product may have or have not specific features. It may be designed for a specific customer segment. It may be cheap or premium, simple or sophisticated. The results of your research will shape these decisions.

In our future chapters we will introduce the concept of product genotype, which together with the customer genotype will help us design a product that matches the customer needs with a feasible and valuable solution.


Key messages are what you tell customers about your product and value proposition. In fact, these messages are rarely fully honest. Promotion always overpromises and extols the virtues of the real product. However these messages contain signals that may or may not trigger the customer’s attention and interest. Thus, experimenting and analyzing what messages resonate with the customers in specific channels are the rightful ways to evaluate demand.


As stated above, the real sales and profits are the most important sign of real demand. Profits however show the retrospective information for the previously designed and delivered products. For a brand new product or a significant product improvement the other methods of analysis are required.

Problem analysis

The main problems in the demand research:

  • People don’t always say what they really need, so the survey results are usually biased
  • Customer interviews require a lot of work and a large and representative sample of customers to make statistical sense
  • Competitors protect their trade secrets that greatly complicates the analysis especially in the B2B market
  • Customer acquisition cost may be too high to ensure statistically significant results of your market research
  • Behavior/sales analysis for the existing products is retrospective and requires very fast release and sales cycles which may be feasible for digital products, but are hard to implement in many other industries

There are basically two factors that determine the most effective way of understanding the demand: cost of customer acquisition and cost of product release.

Low cost of customer acquisition and low cost of product release. This is a great opportunity for experimentation and fast build-measure-learn cycles. You don’t need to worry about the dissatisfied customers because you can easily acquire several hundreds of fresh customers and run a lot of experiments live with product features and promotion messages, observe how customers react to those changes and quickly ship a new product release with the most successful solutions.

Example: digital products for B2C markets that have driven a venture boom in the Silicon valley.

High cost of customer acquisition and low cost of product release. When the cost of customer acquisition is high, you can not simply run experiments with the unprepared real users and see how they behave. Most experiments fail, so most of the users will be frustrated and leave, while you have no way of easily acquiring a couple of hundred fresh users. However, the low cost of product release gives you the privilege of experimenting with a small group of loyal users and very fast prototyping, beta testing, and getting feedback from real experience. Human-assisted testing sessions allow for compensating the negative situations if they occur to the user. The problem however is that the small number of users is underrepresented for making generalized conclusions about the demand. Thus many products in this market are custom tailored to a specific niche and a customer segment.

Example: B2B software is usually simpler to update than physical products, but hard to sell. Quite frequently a B2B product is co-created with several key customers.

Low cost of customer acquisition and high cost of product release. Quite frequently in this quadrant the demand is very much known, but the creation of a new product is too expensive because of technological challenges. Who would not want a flying car? Who would not want an unlimited battery? Who would not want a cure for all diseases? There are billions of customers for such products, but it is hard to develop a cheap enough and functional product that can solve this strong or even wicked problem. Thus the main research activity here is not in collecting data about demand in general, but identifying the customer segments that are ready to buy and try unfinished products. That is products that are not yet complete for the mass market, but have limitations that satisfy the limited segment of users. These customers are ready to pay a much higher price for the product, are ready to take risks, or they volunteer to have early access to the product with strong non-disclosure conditions. Most research budgets for this type of products are not spent on the demand and problem analysis, but on the solution design and experimentation with the underlying technologies.

Example: Clinical trials with new drugs for a very limited group of eligible patients, the first cars being expensive hand-made limited edition products for a small group of wealthy customers before they became a commodity, focus groups and closed film screenings for TV shows, and early access programs for selected certified partners.

High cost of customer acquisition and high cost of product release. This type of business is typical for large industrial products with a very long cycle of design, development, and testing. In such situations the customers are very deeply involved in the process. Instead of the demand research, the architects from both sides, the customer and the product vendor, work together to specify the requirements. The number of customers is small as well as the number of contracts, however the checks may be very high.

Examples: complex industrial equipment, jet engines, aerospace components, and bridge and skyscraper construction projects.

Depending on the type of business of your quest giver, different methods of demand research will be used. As we see, if the customer acquisition cost is high, it is almost the only way to have a small number of customers co-creating the product and sharing their ideas. When the customer acquisition cost is low, you may rely more on statistical methods and take care about having representative groups of customers and unbiased conclusions.

Demand research leveling guide.

Leveling guide — Decision/distribution chain

In any case, there are some general patterns that can help you understand your customer needs and achieve great product-market fit.

If you take part in designing and building a new product from scratch:

Create Placeholders of all elements of the research landscape. And spend some time on exploring the customers segments, run surveys, observe the competitors, so that you have a preliminary Quick & Dirtyunderstanding of what customers need. Your list of needs may contain 20–30 main needs validated by 5–20 customers and market experts.

Collect more information and run more interviews, surveys, expert sessions and competitor analysis to upgrade the list of needs to the Good Enough level (40–50 needs with statistically significant conclusions, 40+ interviews with clients, 400+ of remote surveys for mass market products). Now you have a better understanding of real needs so you may start not just collecting data, but also selling a non-existing product. Use the exact customer wordplay to describe what product does and address the customers in the relevant channels. Observe what messages resonate best with the clients, these are signs of a need. For those customers who are ready to buy the product immediately, you may organize a waiting list, so you don’t lose your face because the product is not ready yet.

Having all the previous steps completed, you can now create an early Quick and Dirty product prototype or a minimal viable product, that reflects the most important customer needs and give into the hands of the limited cohort of customers. The drawbacks of the product should be compensated by the human-assisted testing process. In the best case scenario, you should get paid by these first users which is a good sign of the real demand. The profitability and unit economy of the product may be far from perfect, but you are still in the Quick and Dirty mode, so it is ok.

If you’re researching demand for an existing product with the goal of increasing sales, skip phases 1–3 and go to the next phase.

Even if the product has some revenues, again it is important to check that it solves the real needs of the customer. The market may be evolving and some of the previously identified needs can become irrelevant, competitors also improve their products. So three things that you may immediately start with is the deeper analysis of competitors and segments. Spend more research effort on them to obtain Good Enough vision for these areas. Also analyze the available sales and usage statistics, customer requests and complaints.

Now it is time to focus and upgrade the product. Reprioritize the roadmap to reflect the identified unresolved critical needs, shorten the release cycles and add usage analytics capabilities if possible.

With the newly updated product, reengineer the marketing messages to reflect real needs and start promotion to the most relevant channels. Close the loop and integrate the sales and usage statistics into systematic updates of the customer demand. Repeat and improve all elements of the cycle.