At a planning meeting of an Almaty analytics reporting service, people are arguing about where small retail chains will choose a supplier a few years from now. Today some clients come directly. Zhanar wants to prepare for intermediary catalogues, while Altynai, who runs marketing, is sure digital assistants will soon be doing the comparing. Talgat, the owner, listens to both and still does not know the one thing he needs: where to move the budget.
He needs a decision for today: tidy up the product information, test a new way of acquiring clients, or invest in an expensive integration. Scenarios help split this into three parts: what to do now, what to test with an experiment first, and under what conditions it is worth spending more.
Possible ways to choose a supplier
Conditions to examine, not a country forecast
Direct route
A clear offer and direct contacts
How does the customer find and compare companies?
Intermediary
Accurate information, access and cost
What data and responsibilities does each party receive?
Mixed route
Consistent promises and enquiry history
Who owns handover and error correction?
Scenarios can coexist. The company keeps its own source of offer data.
Separate an observation from its explanation
A signal might be a client's story: first they compared suppliers in a new service, then contacted the company they picked. That is an observation about one purchase. The hypothesis is broader: this way of choosing could become important for the whole segment. Testing it takes other cases and an understanding of what role the service played in the decision.
A scenario describes possible conditions: who helps the buyer choose, who keeps the contact, what information reaches the seller and what has to be paid for. A vision describes the role the company wants for itself. Talgat may want to sell directly, and buyers may still go to intermediaries.
The Government Office for Science Futures Toolkit treats scenarios as possible futures and suggests testing decisions under different conditions. The same approach works for our service. The options in this lesson are conditions to work through, not a forecast for Kazakhstan or any other country.
In an observation log, record the source, the date, the context, the explanation and another possible explanation. A client might have opened a catalogue once because a friend recommended it. That is a very different thing from a habit of starting every choice there. What to do next depends on which explanation holds up.
Describe differences that change the economics
For the service, take three possible ways of buying.
| Condition | What matters to the company | What question stays open |
|---|---|---|
| The buyer chooses a supplier directly | A clear offer, its own contacts and a way to show the result | How the buyer finds and compares suitable companies |
| An intermediary runs the comparison | Accurate information, terms of access to demand and acquisition cost | What data and obligations each party gets |
| Choice and service happen in different places | The same promises everywhere and a preserved enquiry history | Who owns the client handover and fixing mistakes |
A digital assistant could turn out to be part of any of these options. But the fact that an assistant has appeared in the chain does not tell you who gets the money, who passes on the order or who answers for the result. Describe those relationships before you judge the channel.
Scenarios can exist at the same time. The service can serve direct clients and get some enquiries through an intermediary. So the task is not to guess one future. What matters is seeing where costs, access to the buyer and the ability to keep a promise change.
Put one information set to a real test
The service's offer includes two supported accounting systems, a seven-day pilot and an agreed report scope. The working description needs verified system names and versions, pilot start conditions, limits and a person who keeps the description up to date. A stable offer ID will help match versions on the website, in the CRM and in a catalogue export.
This is needed now: Zhanar's managers must not promise one client different terms. In direct sales the information serves the website and the manager. If an intermediary appears, it is easier to pass on and reconcile. On a mixed route, it is easier to tell which version of the offer the buyer saw.
There is no need to build one format for every future intermediary: each one's requirements are checked when it shows up. But the service should keep its own source of data, even if the same information also sits in the fields of someone else's interface.
Estimate the work before you start. Pilot preparation is budgeted at eight hours of Yernar's time (he runs delivery) at an internal rate of 5,000 ₸, 40,000 ₸ in total. All the numbers in this lesson are made up. In that time he tidies up a limited offer and checks an export. If his pay does not change, that 40,000 ₸ does not become an additional cash expense.
The fixed 6,800,000 ₸ in the service's financial model already includes the whole team's payroll, rent and shared services. Yernar's salary is already part of that sum, so his 40,000 ₸ is not subtracted a second time. For the next calculation, assume something different: the service buys extra outside work for the pilot, 40,000 ₸ to prepare the information and 60,000 ₸ for the integration. Additional costs then come to 100,000 ₸.
Calculate a limited test
The monthly package costs 60,000 ₸, variable costs per delivered month are 20,000 ₸, and contribution before acquisition and fixed costs is 40,000 ₸. The new intermediary charges 15,000 ₸ for each new buyer after their first paid and delivered month, and this is not the rate of any real platform.
After the fee, 25,000 ₸ remains per buyer. Covering the 100,000 ₸ of preparation and integration takes four such buyers. This calculation neither covers nor removes the service's fixed costs: they stay in the overall P&L.
The pilot plan has six new buyers. If each one pays for and receives the first month, revenue will be 360,000 ₸, delivery will cost 120,000 ₸ and the intermediary's fees 90,000 ₸. That leaves 150,000 ₸, of which 100,000 ₸ covers preparation and integration. On plan, the pilot makes 50,000 ₸ before shared fixed costs and tax.
To compare this route with the direct channel, count the costs of both within the same boundaries. In the month reviewed earlier, 200,000 ₸ of advertising brought in ten new companies, an ad-only CAC of 20,000 ₸. The 15,000 ₸ fee looks cheaper, but the new route has preparation costs, and nobody knows yet how much work these clients will take. Two rates are not enough for a comparison. And do not add returning clients to the new ones to make CAC look lower; count a company that came through several touchpoints once.
Agree what would justify continuing
Before the pilot, agree which enquiries belong to the chosen segment, how a new client is defined and when the fee counts as earned. Match IDs against current clients and open deals. Otherwise the report will credit the intermediary with a purchase Zhanar was already preparing directly.
Set a limit on additional spending and on the amount of work the team is willing to take on. Follow the path to payment and delivered service, the reasons for refusals, the support costs and whether the promise matched the result. Offer-card views will help find breakdowns, but they will not replace this evidence.
It is worth continuing to invest if suitable enquiries have started coming in, the cost per delivered order is acceptable and quality has not dropped. If there is too little data, a sensible decision is to keep observing on a limited scale. If the conditions clearly do not work, stopping the pilot saves money for other tasks.
Read news through the decision you need to make
A feature announcement, a demo and a purchase path that works reliably are three different grounds for action. For a chosen platform you need its current terms, whether the scenario you need is available, and its limits. For a loud claim about an audience, what matters is the research method and whether the sample looks like your segment. Until you have checked, keep such information in the assumptions column.
Another market's experience can suggest a question, but it will not answer for you about language, payment, trust and service in Kazakhstan. There is no need to collect every technology headline. It is enough to know which new evidence would change your decision: buyers have started looking for suppliers differently, access to them has become more expensive, data has become unavailable, or the conditions for delivering the service have changed.
Exercise: twelve mentions and three customers
In 12 of 100 interviews, people mentioned a new comparison service. Talgat proposes moving 12% of the 700,000 ₸ budget there, that is 84,000 ₸. Instead, the team ran the pilot from this lesson, with additional costs of 100,000 ₸. By the review date, only three new buyers had paid for and received their first month; three more are still a forecast. Work out what you can say about demand and about the result.
The service came up in 12% of these interviews. That is a share of one group of interviewees: it does not tell you the market share, and there is no basis for moving the budget in the same proportion. You need to rebuild how people actually chose a supplier and understand the intermediary's role. The three completed orders brought 180,000 ₸ of revenue, with 60,000 ₸ spent on delivery and 45,000 ₸ on fees. A contribution of 75,000 ₸ does not yet cover the additional 100,000 ₸, so the result is minus 25,000 ₸ before shared fixed costs and tax. The remaining three could change the outcome once they pay and receive the service, but it is too early to record them as income received.