Market research for a small business in Egypt: estimating demand from official data and fieldwork

Many weak feasibility studies are not weak in the arithmetic. They are weak on the first line: expected sales. That figure does not come from the cost sheet or from the founder's ambition. It comes from market research for a small business: knowing who your customer is, how many of them live or work where you will operate, how much they spend, and who is competing for them. If that number is weak, everything built on it, from break-even to payback, is just as weak.

This article explains market research in plain terms: the difference between desk and field research, Egypt's official data sources and their limits, how to size a market two ways and compare the results, and a fully worked illustrative example for a small bakery in a residential neighbourhood. It also covers how to choose a survey sample size and how the result turns into the inputs of a feasibility study.

What is market research, and why does a small business need it?

At its simplest, market research is the collection and analysis of data about a business's target market¹. The target market is the group within the total market that you will serve, chosen for its ability and willingness to buy¹. In everyday terms: not everyone in your city is your customer, only the people who need what you sell, can pay for it and can reach you.

In a feasibility study, the market analysis identifies competitors and puts a number on target customers². A market can be described by its size, structure, growth prospects, trends and sales potential². This is why market research matters more, not less, for a small business. A large company can absorb a wrong demand estimate. The owner of a shop or workshop who puts their savings into it can lose everything on one untested assumption.

Three questions it must answer

1. Who exactly is the customer? Households in a particular area, students, small companies, car owners. A target market can be segmented by geography, customer attributes or product type².
2. How big is demand? How many potential customers are within your reach, how often each one buys and how much they spend each time.
3. Who serves them today, and at what price? Direct competitors, the substitutes customers use and the prices they are used to. This determines the share you can take and the price you can charge.

Desk research and field research: what is the difference?

Every market study uses two kinds of data. Secondary data (desk research) is existing data collected by someone else¹ for a purpose other than your project³, such as population statistics or counts of establishments. Primary data (field research) is new data you collect to answer a specific question¹, with your current project in mind³, such as counting passers-by outside a site or surveying residents of a neighbourhood.

The practical rule: start with desk research because it is cheap, fast and gives you the frame, then move to field research because only that tells you what happens on the street where you will open. The table below summarises the most common tools:

Observational research is a field method in its own right: data collected by watching consumers and recording what they do³. Its strength is that it measures actual behaviour rather than what people say about themselves. A focus group is a small group, typically 8 to 12 people, asked several questions by a moderator³.

Official data sources in Egypt and how to read them

The state collects a lot of data that can help a small-business owner, provided you know what each source measures and which period it covers.

Population and the census

Egypt's resident population reached 109.5 million in September 2026, according to preliminary data from CAPMAS, the national statistics agency⁴. The national figure helps you understand overall scale, but it is of little use to a business serving a single neighbourhood. What matters more is the general census of population, housing and establishments. The 2017 census collected data on households and individuals, on buildings, and on establishments, including their activity, legal form and number of employees, using geographic maps down to building level⁵. In May 2026, final preparations were under way for the 2027 general census of population, housing and establishments⁶. In practice, the latest detailed published population and housing data for your area may date from 2017, so treat it as a starting point and correct it with what you see on the ground, such as new buildings and areas that have filled up since.

Establishment counts: a signal of competition

According to CAPMAS data, Egypt had about 3.9 million establishments in 2023. Cairo led the governorates with about 484.6 thousand, followed by Giza with 379.1 thousand, Dakahlia with 301.8 thousand and Sharqia with 284.5 thousand⁷. These figures will not tell you how many competitors are on your street, but they help you compare governorates and ask the right question: is this activity already crowded where I am? For a wider reading of what national data says about small-business opportunities, see small business ideas in Egypt based on data.

Income and expenditure

CAPMAS has run its household income, expenditure and consumption survey every two years since 2008⁸. In August 2025 the agency said that multidimensional poverty indicators from the 2021/2022 survey had been published in November 2024, and that the 2023/2024 report was being prepared and expected in mid-October 2025⁸. We could not confirm at the time of writing whether the 2023/2024 report has been released, so check for the latest edition yourself on the day you write your study, and keep two things in mind. The figures are averages for all households, not for your segment. And their value in pounds ages fast: annual urban inflation was 13.9% in September 2026, after 14.5% in August⁹, so a spending figure that is three or four years old cannot be used as it is to estimate today's sales. Use these surveys for priorities and proportions, and take prices from the market now.

Online search trends

Google Trends shows how interest in a product or service changes over time and across regions. Read it for what it is: it uses a sample of searches, its figures are scaled from 0 to 100 to compare relative popularity, and low-volume terms show as zero¹⁰. It is useful for seasonality and direction, not for estimating the number of customers.

Market size: total, serviceable and obtainable

When you estimate demand you will meet three levels, and it is important not to mix them up:

• Total available market (TAM): the total perceived demand for the product or service in the marketplace¹. For a local business, you can treat it as everyone who buys this type of product in the city or the wider district.
• Serviceable available market (SAM): the part of the market your business can serve given its products, services and location¹. For a shop in a residential area, it is usually the residents who can reach you on foot or by quick delivery.
• Obtainable share: the part of the serviceable market you expect to win given the competition. Projecting market share can be a subjective estimate², so your study should explain why you chose the percentage and test it a second way.

The usual mistake is to jump straight from the total market to sales: "the bakery market in Egypt is worth billions, and if we take 1% of it..." That sentence says nothing about your business, because 1% of a market you cannot reach is not sales.

Two ways to estimate demand in small-business market research: top-down and bottom-up

Top-down: you start from the big picture and broad economic factors to estimate an industry's potential¹, then narrow it: number of households, share who buy, average spending, the part you can serve, then your share.

Bottom-up: you start on the ground, counting local market sizes and competitors' numbers of customers¹, then add up the local markets and multiply by the average annual spend on the product¹. For a single shop, that means in practice: how many people pass your site, how many of them will walk in and buy, and how much they will pay on average. You can also look at similar businesses in similar locations to gauge how yours might perform².

Why both? Because each method goes wrong in a different way. Top-down tends to be optimistic because it starts with a large number and applies assumed percentages to it. Bottom-up depends on a field count that may not be representative if it was done on one day or at one hour. If the two results are close, you are on firmer ground. If they are far apart, find the wrong assumption before going further.

Worked illustrative example: market research for a small bakery

This is an illustrative example. All its figures are assumptions made to explain the method; they are not data about a real neighbourhood or averages for bakeries. A founder is considering a small bakery (bread rolls, pastries and simple sweets) in a residential neighbourhood, on a side street with pedestrian traffic.

Method 1: top-down

• Households in the neighbourhood: 15,000.
• Share of households buying these products from a bakery every week: 60%, or 9,000 households. This is an assumption to be tested with the survey.
• Average spending per buying household: EGP 150 a week, or 150 × 52 = EGP 7,800 a year.
• Total market in the neighbourhood: 9,000 × 7,800 = EGP 70,200,000 a year, or EGP 5,850,000 a month.
• Serviceable market: 6,000 households live within walking distance or a few minutes' delivery. 60% of them buy, or 3,600 households, so the serviceable market is 3,600 × 7,800 = EGP 28,080,000 a year, or EGP 2,340,000 a month.
• Obtainable share: four bakeries already operate in the same area. If five bakeries split the market equally, each would have 20%, but a new bakery does not start with the share of established ones, so we assumed 8% in year one: 28,080,000 × 8% = EGP 2,246,400 a year, or EGP 187,200 a month.

Method 2: bottom-up

The founder counted passers-by outside the site on three different days, in the morning, midday and evening: 1,400 people on an ordinary weekday, 1,800 on Thursday and 1,700 on Friday. Because a month has five ordinary weekdays for every Thursday and Friday, they took a weighted average: (5 × 1,400 + 1,800 + 1,700) ÷ 7 = 1,500 people a day. If Saturday traffic differs in your area, count it separately and give it a weight of one day in seven. A simple average of the three days (about 1,633) would give the busy days more weight than they really have and overstate sales. They visited the four bakeries to record prices and what a typical customer buys. Then they assumed:

• Share of passers-by who walk in and buy: 3%, or 1,500 × 3% = 45 customers a day.
• Average in-store ticket: EGP 90, or 45 × 90 = EGP 4,050 a day.
• Delivery orders by phone and WhatsApp: 10 a day at an average of EGP 140, or EGP 1,400.
• Daily sales: 4,050 + 1,400 = EGP 5,450, and over 30 days: EGP 163,500 a month.

First comparison

The bottom-up estimate (EGP 163,500) came in EGP 23,700, or about 12.7%, below the top-down obtainable share (EGP 187,200). The gap is not large, but the riskiest number in method 1, the share of households who buy (60%), is still a guess. So we do not adopt either figure until that share has been measured.

Correcting the assumption with a survey

The founder surveyed a random sample of 100 households within the serviceable market, and 48 said they buy these products from a bakery every week. The true share therefore probably lies between about 38% and 58% (the survey section below shows how this range is calculated), and even the upper bound is below the 60% we assumed. We re-ran method 1 at 48%: 6,000 × 48% = 2,880 households, a serviceable market of 2,880 × 7,800 = EGP 22,464,000 a year (EGP 1,872,000 a month), and an obtainable share at 8% of EGP 149,760 a month. The same share also cuts the total market in the neighbourhood to 15,000 × 48% = 7,200 households, or 7,200 × 7,800 = EGP 56,160,000 a year.

Now the field estimate (EGP 163,500) is EGP 13,740, or about 9.2%, above the top-down figure (EGP 149,760). The cautious rule is to adopt the lower figure in each scenario and treat the higher one as a ceiling, not a target. To build the scenarios we used the ends of the survey range in method 1, and changed the share of passers-by who buy and the number of delivery orders in method 2:

Look at the last row: the optimistic field count (EGP 220,800) is EGP 39,840 above the method 1 ceiling, because it implies the new bakery would take more of the serviceable market than the 8% share we assumed. That is not impossible, but it needs a clear reason, such as a product competitors do not offer, a better price or a stronger location; otherwise the ceiling is the honest figure.

Field research step by step

1. Foot-traffic counts and observation

Choose at least three days that include a weekday and a weekend day (count Saturday separately if its traffic differs), and split each day into periods (morning, midday, evening). As a rule of thumb, count passers-by for 15 minutes in each hour and multiply by 4, then take an average weighted by the number of each kind of day in the month, as in the example, and record how many people enter nearby competitors. Watch what people carry out too: it is the closest you can get to measuring actual purchases without asking anyone.

2. Competitor visits

Visit each competitor at least twice at different times, and buy something. Record prices, pack sizes, opening hours, waiting time, payment and delivery options. Do not write "competition is weak" in your study without evidence; write what you saw: how many competitors, their prices and what they lack. Then use those prices to set yours with the method in how to price a product.

3. Surveys and sample size

A sample is the portion of the whole population included in the research³, and the sample frame defines who is included in the sample and who is not³. In the bakery example, the sample frame is households living within the serviceable market, not your friends on social media.

To estimate a proportion (such as the share of households who buy every week), sample size is z squared × p × (1 − p) ÷ the margin of error squared¹¹. If you do not know the expected proportion, using 0.5 gives the largest required sample¹¹, so it is the safe choice. At 95% confidence, z is 1.96:

• Margin of error of 10 percentage points: 1.96² × 0.5 × 0.5 ÷ 0.10² = 96.04, rounded up to 97 households.
• Margin of error of 5 percentage points: 1.96² × 0.5 × 0.5 ÷ 0.05² = 384.16, rounded up to 385 households.

In the textbook's own example, a 3-point margin at 90% confidence (z = 1.645) gave 751.7, so the sample size became 752¹¹. Note that halving the margin of error roughly quadruples the sample. For a small business, about 100 well-chosen households from your area are far more useful than 400 answers from outside it.

How do you read the result? If 48 of 100 households say they buy these products every week, the margin of error at 95% is 1.96 × the square root of (0.48 × 0.52 ÷ 100) = about 9.8 percentage points, so the true share probably lies between about 38% and 58%. This is exactly what we used in the example to correct the 60% assumption. Note that these calculations assume a random sample from the sample frame. Asking whoever you meet at one door or at one hour can give a biased result however many answers you collect, so spread your questions across different streets and times.

4. Interviews: why people buy or refuse

A survey gives you proportions; an interview gives you reasons. As a rule of thumb, talk to 15 to 20 potential customers about the last time they bought this product: where, for how much and why there. Ask about the past, not the future, because "would you buy from me?" usually gets a polite yes. You will find how to run these interviews and test demand with a deposit or pre-order in how to test your business idea before spending your capital.

Common market research mistakes

1. Starting with the national market figure and assuming a percentage of it, instead of starting with the customers you can actually reach.
2. Using old data as if it were today's, especially spending figures in pounds when inflation is high, or population counts for an area that has changed a lot.
3. Asking friends and family instead of a sample from the business's catchment, so the results are polite and unrepresentative.
4. Confusing purchase intent with purchases. People who say "I would buy" in a survey may not, so test demand with pre-orders or deposits whenever you can.
5. Counting passers-by on one day or only at peak hour, then multiplying by 30.
6. Ignoring indirect competitors, such as what households make at home or buy at the supermarket.
7. Assuming a market share with no explanation. Write down why you chose it and compare it with a second, bottom-up estimate.
8. Starting at full sales from month one. Market research estimates sales once the business has settled; getting there takes months of ramp-up.

From market research to feasibility study figures

Market research is not a separate chapter of the study; it is where the most important inputs come from:

• Expected volume: in the example, adopted realistic sales of EGP 149,760 a month, or about EGP 4,992 a day. At the same sales mix as the field count (about EGP 99 per sale across shop and delivery), that is about 50 sales a day: about 41 in store and about 9 delivery orders.
• Price and average ticket: from competitor visits and the survey; in the example, EGP 90 in store and EGP 140 for delivery.
• Scenarios: pessimistic, realistic and optimistic, taking the lower of the two estimates at the ends of the survey range, as in the example table.
• Sales ramp-up: do not start at settled sales in month one; assume a gradual climb.

You then compare expected sales with costs to find break-even and the capital you need, as explained in break-even and working capital. If, for example, monthly break-even sales are above the pessimistic scenario, you know before you start that the safety margin is thin. The full picture of what a study contains is in the feasibility study guide for Egypt 2026.

How Kleverz helps

Kleverz is an AI-powered platform for feasibility studies. If you are still at the idea stage, the free idea check takes about 5 minutes and gives you a verdict (proceed, adjust, validate first or avoid), a Kleverz Score out of 100 across five dimensions, the top 3 risks and a first step to take in the next 48 hours.

Once you have your market figures, the detailed study is built on the numbers you enter: price, direct cost, expected volume, team, city and capital. It calculates start-up costs and fixed monthly costs including the employer's social-insurance share, a month-by-month cash flow for the first 12 months and a yearly summary over three years, monthly break-even revenue and the units needed per product or service, the operating break-even month, the payback month, the deepest cash need and a recommended capital that adds a two-month reserve, with optimistic, realistic and pessimistic scenarios, the main risks, indicators to track and a week-by-week plan for the first month. It is delivered within 24 hours of payment confirmation, in Arabic or English. Because the study is built on the expected volume you enter, good market research is what keeps its results close to reality. When you decide to go ahead, Maher Consulting Firm, the human consulting arm, handles company formation.

Frequently asked questions

Sources

1. OpenStax. Entrepreneurship — 8.2 Market Research, Market Opportunity Recognition, and Target Market. Checked 26 October 2026.
2. OpenStax. Entrepreneurship — 11.3 Conducting a Feasibility Analysis. Checked 26 October 2026.
3. OpenStax. Principles of Marketing — Chapter 6 Key Terms. Checked 26 October 2026.
4. Parlmany. 1.4 مليون نسمة جديدة خلال عام.. سكان مصر بالداخل يرتفعون من 108.1 مليون فى سبتمبر 2025 إلى 109.5 مليون فى سبتمبر 2026 (in Arabic) (2026-10-06) — secondary source. Checked 26 October 2026.
5. CAPMAS. Arab Republic of Egypt - General Census for Population, Housing and Establishments 2017. Checked 26 October 2026.
6. Al-Ahram Gate. وزير التخطيط ورئيس «التعبئة والإحصاء» يبحثان الاستعدادات النهائية للتعداد العام 2027 (in Arabic) (2026-05-17) — secondary source. Checked 26 October 2026.
7. Banker. 1.7 مليون عامل إضافي.. ارتفاع أعداد المشتغلين في مصر إلى 15.2 مليون عام 2023 (in Arabic) (2025-11-28) — secondary source. Checked 26 October 2026.
8. Al-Masdar. «الإحصاء» يؤكد صدور بحث الدخل والإنفاق للعام المالي 2021-2022 (in Arabic) (2025-08-04) — secondary source. Checked 26 October 2026.
9. Enterprise. Urban inflation eased to 13.9% in September, even as monthly prices continue to pick up (2026-10-11) — secondary source; Zawya. Egypt's annual urban inflation eases to 13.9% in September (2026-10-12) — secondary source. Checked 26 October 2026.
10. Google. FAQ about Google Trends data — Trends Help. Checked 26 October 2026.
11. OpenStax. Introductory Business Statistics — 8.4 Calculating the Sample Size n: Continuous and Binary Random Variables. Checked 26 October 2026.