Every advertiser in Ethiopia plans against the same figure. Satellite operators, industry bodies, and media owners repeat it with the confidence of a settled fact: approximately 17 million TV-accessing households make up the Ethiopian broadcast universe (I have even encountered recent data that states 18.5 million households). It is the denominator for every rating point, the foundation of every reach claim, and the number on which hundreds of millions of birr in advertising is allocated each year. Taken at face value, it means Ethiopian television reaches more than half the country.

The figure is not fabricated. It is a count of a real thing. The more productive question is what that thing actually is, how it relates to the commercial decisions being made against it, and whether it is the right denominator for the right objective. When you work through the data carefully, you find not one planning universe but four, each with different commercial implications and each requiring a different denominator. Using the wrong one for the wrong objective is not imprecision. It is a misalignment between investment and outcome.

Getting the denominator right

The 2024-25 Ethiopia Demographic and Health Survey, published by the Ethiopian Statistical Service in January 2026, places average household size at approximately 4.73 people. That number is the foundation of every estimate that follows.

Working from a population of approximately 138.9 million:

138,900,000 / 4.73 ≈ 29,366,809 total households

You read this as: Ethiopia has roughly 29 million distinct households when you divide the national population by how many people typically share a home. The EDHS itself cites 21.9 to 22.5 million, which is the more grounded figure, derived from actual household enumeration rather than population math. The EDHS sampling, however, was conducted over a reference period that predates the 2026 population estimates used here, and given Ethiopia’s urbanisation rate of 4.9% per year, the true household count today likely sits somewhat above what the survey recorded. A reasonable working range is 22 to 25 million total households, with the number trending upward as urban household formation continues. This analysis uses 29.3 million as a ceiling figure precisely because it works against the argument being made here. The lower the true household count, the smaller the commercially addressable audience as a share of the total, which only sharpens the case for planning against the right denominator rather than the largest available one.

Ethiopia’s urban population in 2026 is estimated at approximately 31.8 million, representing 22.9% of the total.

Dividing that urban population by an urban average household size of 3.9:

31,800,000 / 3.9 ≈ 8,153,846 urban households

A separate calculation using the national household total and urbanisation rate produces 6.7 million. Both approaches land urban households in the same general range, and seven million is a reasonable working figure for the analysis that follows.

Broadcaster Takeaway: The 17 million figure, or 18.5 million depending on the source, is not a number to hide from. It is a number to explain. Every time a broadcaster presents it as a single commercially active audience without qualification, they are one client audit away from a credibility problem. Get ahead of it. Know what the number actually measures before someone else defines it for you.

Why the electricity data complicates everything

If 17 million households genuinely have television access, that represents between 58% and 78% of all Ethiopian households, depending on which total household figure you accept.

17,000,000 / 29,366,809 = 57.9% 17,000,000 / 21,900,000 = 77.6%

Both figures sit uncomfortably against the electricity data, though the electricity picture requires two distinctions to read correctly. The first is connection: whether a household has any form of electricity access at all, including solar and micro-grid sources. The second is reliability: whether that connection delivers enough consistent power to run a television for several hours on a typical evening. These are different thresholds and they produce different numbers.

On connection, preliminary findings from the 2025 World Bank Multi-Tier Framework Survey place Ethiopia’s Tier 1+ access rate at 44%, meaning roughly 71 million people, primarily in rural and peri-urban areas, remain without even a basic electricity supply. On reliability, the picture is sharper at the extremes: 93% of households in Addis Ababa are connected to the grid, while only 22% of households nationwide have access to a supply consistent enough for regular appliance use.

Reliability, not mere connection, is the gating variable for television viewing. A household with a solar lantern sufficient to charge a mobile phone is counted in the Tier 1+ figure. A household that can run a television for three hours on a weekday evening is a categorically different thing. The 17 million figure almost certainly includes decoder registrations in households without reliable power, community viewing points counted as discrete access units, and sets owned but infrequently powered due to supply constraints. None of this makes the number dishonest. It makes it a measure of access opportunity rather than active commercial viewership, and those are different denominators for different purposes.

Agency Takeaway: Planning cost per thousand against 17 million households for an FMCG brand means your efficiency calculation is spread across a universe three to four times larger than the audience that can actually act on the advertising. The budget is not too large. It is allocated against the wrong base. Correcting the denominator does not mean spending less. It means every birr works harder against the households that are genuinely in a purchase cycle.

What living standards actually measure

Before going further, it is worth explaining the framework this analysis rests on, because it is not widely understood outside research circles, and because it answers a question that tends to surface at this point in the argument: why not just use income?

An LSM score is built from observable variables: the material a home’s walls are made of, whether it has a flush toilet or a pit latrine, whether it has electricity, what cooking fuel is used, what level of education the household head completed, and what assets the household owns. Each variable is scored and weighted. The total score turns out to correlate more reliably with actual consumption behaviour, brand purchasing, and media access than income alone. You cannot hide a corrugated iron roof or the absence of a water connection. LSM is honest data precisely because it is externally verifiable in a way that income declarations are not.

Think of it less like a bank statement and more like a walk through someone’s home. Two households can report similar incomes and live entirely different lives, because one is in central Addis Ababa with a television, a gas cooker, and a functioning water connection, while the other is in a peri-urban settlement with intermittent electricity and a shared outdoor latrine. An LSM score captures that difference. An income figure often does not.

This distinction matters beyond data hygiene. Income tells you what a household can afford to buy today. LSM tells you what it has the capacity to consume every day. Those are not the same thing, and the gap between them is where most media plans quietly go wrong.

Consider a household earning a reasonable daily wage but living off the grid in a peri-urban settlement. That household can technically buy a sachet of shampoo, a bouillon cube, or a snack. It is buying for today, in single-use quantities, driven by whatever cash is in pocket at that moment. It cannot buy a two-kilogram pack of detergent, bulk cooking oil, or fresh dairy, not because it lacks the income on a given day, but because it lacks the electricity, refrigeration, and storage conditions that make those purchases possible and rational. Its income is real. Its FMCG consumption potential for a large portion of branded categories is close to zero.

This is what separates what we might call the sachet consumer from the pantry consumer. The sachet consumer buys for today. Their purchasing is real, frequent, and worth something to certain categories, particularly single-use personal care products, small-format snacks, and basic consumables. But their spending is driven by daily income availability rather than deliberate brand choice, and it is highly vulnerable to income shocks, seasonal employment changes, and inflation. The pantry consumer buys for the month. Their household has the physical conditions, consistent electricity, storage space, and income stability, that make regular branded purchases across multiple categories a sustained behaviour rather than an opportunistic one. For an FMCG brand building market share rather than just chasing transactions, the pantry consumer is the commercially relevant target. LSM scores identify pantry consumers with a reliability that income brackets cannot match.

The framework used in this analysis adapts the LSM methodology to the Ethiopian context using the Ethiopia Socioeconomic Survey from 2021 and 2022, the Ethiopia Demographic and Health Survey from 2019, and World Bank poverty assessment data from the same period. Together they score Ethiopian households on a 64-point scale across 16 variables. The national median score at the time of the surveys was 11 out of 64.

Curious Analyst Takeaway: The sachet consumer and the pantry consumer can have similar daily incomes and look identical in an income bracket. They are not the same commercial target. The sachet consumer is a real revenue opportunity for single-use and micro-format products. The pantry consumer is the foundation of branded market share across categories. LSM separates them. Income data does not. That is the practical reason the framework exists.

The living standards foundation underneath the audience

A household at score 11 has mud or wood walls, an earth floor, a shared pit latrine or open defecation, firewood or charcoal for cooking, no reliable electricity, and no television. This is the median Ethiopian household. More than half the country sits at or below this point. It is not a marginal or edge-case population. It is the middle of the distribution.

The surveys are between four and seven years old at the time of writing, and Ethiopia is not a static country. It is urbanising at approximately 4.9% per year, one of the fastest urban growth rates in Sub-Saharan Africa, adding roughly 440,000 people to its urban population annually. At an average urban household size of 3.9 people, that translates to approximately 113,000 new urban households forming each year and crossing the threshold from rural living standards into the commercially addressable band.

Compounding that rate across the gap between the oldest survey data and today:

113,000 households/year × 5 years (2021 to 2026) ≈ 565,000 net new urban households

We can postulate that by April 2026, the commercially addressable Grade BC band has grown by somewhere between 450,000 and 565,000 households since the survey baseline, with the range reflecting variation in how quickly newly urbanising households cross the income and asset threshold rather than simply residing in a city. The electrification programme has continued extending grid access into secondary cities over the same period, further lifting households previously excluded from the television-owning distribution on electricity grounds alone. The national median score, if resurveyed today, would likely sit closer to 13 or 14 rather than 11. The break in the distribution is still there. It has moved slightly upward and the population above it has grown. That is the direction of travel, not a revision of the framework.

Urban households continue to cluster between scores 27 and 39. Rural households between 6 and 15. The gap from 15 to 26 remains nearly empty; fewer than 6% of all Ethiopian households score in this range. Living standards in Ethiopia do not improve gradually as you move from rural to urban. They jump.

A household at score 26 is at the urban floor: cement block walls, an indoor toilet connected to a shared sewer, grid electricity that works most days, a mobile phone, a television. Below score 26, almost certainly rural. Above it, almost certainly urban.

One important note on digital before moving into the tiers. National mobile internet user figures for Ethiopia, estimated at 32 to 36 million by end of 2024, can create a misleading impression of broad commercial reach. Digital penetration at the national level is wide but shallow, skewed heavily toward basic data use in lower income bands. What makes digital commercially meaningful is not its national breadth but its concentration in the upper living standards tiers. In the Grade BC band and above, smartphone ownership is near-universal and active data use is a daily behaviour. That concentration is what gives digital its planning value in this market, and it is what the tier analysis below reflects.

Broadcaster Takeaway: Ethiopia is urbanising at 4.9% per year, one of the fastest rates in Sub-Saharan Africa. That means roughly 113,000 new households are crossing into the commercially addressable tier every year. A broadcaster that can show an advertiser where those households are forming, by city, by region, by living standards score, is not selling airtime. They are selling market development intelligence. That is a different conversation at a different level.

Four distinct audiences inside the same broadcast footprint

Rather than one planning universe, the Ethiopian audience is better understood as four tiers: a national awareness audience, a commercially addressable mass, a high-value precision audience, and a formation tier of households not yet commercially active but approaching the threshold. Each has both a television dimension and a digital one. They are not the same audience reached through two channels. They overlap substantially at the top and diverge sharply at the bottom.

The national awareness audience

The first tier is the national awareness audience. These are households with television access in some form, watching intermittently, representing genuine reach for messages requiring broad exposure. A telecommunications company launching a new service, a financial institution trying to build name familiarity, a consumer goods brand entering Ethiopia ahead of its distribution build: all have legitimate use for broad reach, regardless of whether those households are currently making purchase decisions in the relevant category. This audience is the closest to what satellite operators are counting, somewhere between 12 and 17 million access points, with the lower end reflecting regular active viewing and the upper end capturing community and intermittent viewing.

Within this same population, a parallel digital mass audience exists driven largely by Ethio Telecom’s expanding data network and the growth of mobile money across the country. Most of this digital population sits in the same broad awareness tier: reachable, aspiring, not yet in a consistent purchase cycle for most branded categories. Many households that own a decoder also own a smartphone, and for a campaign seeking maximum national impressions the two channels are additive. For a campaign seeking commercial conversion, neither channel alone, nor both together, changes the underlying commercial readiness of a household that has not yet crossed the living standards threshold.

The commercially addressable audience

The second tier is the commercially addressable audience. The LSM framework identifies the Grade BC urban mass segment as households scoring between 26 and 49. These households have permanent urban housing, consistent grid electricity, secondary or tertiary education, a mobile phone, a television, and meaningful access to formal financial services. They are buying branded goods with regularity, have disposable income above the essential spending threshold, and are making active brand choices. They are, in the language introduced earlier, pantry consumers. Their purchasing is deliberate, recurring, and responsive to brand familiarity built over time.

Given that the underlying surveys are several years old and the framework has shifted at the margins, this segment is best expressed as a range: approximately 19 to 21% of all households nationally, yielding a planning range of:

29,366,809 × 0.19 ≈ 5,580,000 households 29,366,809 × 0.21 ≈ 6,167,000 households

You read this as: somewhere between 5.6 and 6.2 million households are in the commercially active urban mass, making brand decisions frequently enough that advertising produces measurable return within a standard planning cycle. Of these, approximately 72 to 77% own televisions, yielding a TV-addressable count of between 4.0 and 4.8 million households. Urban electrification at 94% means this audience’s television is almost certainly on.

The digital layer inside this tier is commercially present but it is not a replacement for broadcast. Grade BC urban households have smartphones and active data connections. Social media, YouTube, and streaming platforms are part of their daily media diet, often alongside rather than instead of television. A combined TV and digital plan reaches this household at higher frequency than either channel alone. But the operative word is combined. Digital does not deliver the same kind of brand exposure. It operates in a different cognitive register, which the section below addresses directly.

Agency Takeaway: The commercially addressable TV audience is between 4.0 and 4.8 million households, not 17 million. That is your FMCG planning universe. Knowing this does not shrink the budget. It sharpens where it goes. A client whose spend is precisely matched to the commercially active tier is getting fuller market coverage of the audience that matters, not a diluted impression across an audience that largely cannot buy.

The high-value precision audience and the premium paradox

The third tier is the high-value precision audience: households scoring 32 and above. Reliable electricity, formal employment, a bank account, household appliances, consistent discretionary spending across multiple branded categories. This is where mental availability compounds, where penetration among high-value buyers is won, and where premium margins are earned.

Expressed as a range to reflect the age and movement of the underlying data:

29,366,809 × 0.11 ≈ 3,230,000 households 29,366,809 × 0.13 ≈ 3,818,000 households

The two estimates converge on approximately 3.2 to 3.8 million households as the core FMCG precision audience, with 3.4 to 3.5 million as the central planning estimate. These are the volume households. They are the engine of category sales across most branded FMCG segments. A can of beer, a family-size bottle of cooking oil, a bar of premium soap: these households buy them regularly, buy them by brand preference, and buy them in quantities that make them the foundation of market share for any serious consumer goods company operating in Ethiopia.

At the narrowest end of this tier sits a smaller, more commercially dense group identified by a single variable: refrigerator ownership.

29,366,809 × 0.01 ≈ 293,668 households

Roughly 294,000 to 330,000 Ethiopian households own a refrigerator. To own a refrigerator, a household needs reliable grid electricity as a non-negotiable precondition, which immediately excludes the vast majority of rural and peri-urban households regardless of income. It needs formal income consistent enough to justify the purchase price and the ongoing running cost. And the presence of a refrigerator fundamentally reshapes the consumption patterns of everyone living in that home.

This is the premium paradox. The 3.4 to 3.5 million Grade BC households are the engine of volume. The 294,000 to 330,000 refrigerator-owning households are the engine of margin. Think of it in terms familiar to any FMCG brand operating in this market. The 3.5 million are the consumers buying a cold Coca-Cola or a BGI beer at a café or a corner kiosk, one unit at a time, in the moment. The 294,000 are the households stocking a two-litre family bottle on a shelf at home, buying in bulk, buying by brand, and restocking with regularity. Both groups matter. They are not the same commercial target, and they do not respond to the same media strategy or the same pack format.

It is worth noting that the home-stocking behaviour of this premium tier extends well beyond refrigerated products. A household with a refrigerator also has the storage conditions, income stability, and shopping habits that support bulk purchasing of ambient goods, large-format detergent, multi-pack snacks, five-litre cooking oil, and branded staples bought for the month rather than the day. The refrigerator is a proxy for an entire consumption infrastructure, not just a signal for cold chain categories.

Kenya, with a national electrification rate of approximately 75%, has a national refrigerator ownership rate of roughly 13%, with urban ownership in Nairobi and Mombasa reaching around 19%. Even in one of East Africa’s most commercially developed urban markets, eight out of ten households do not own a refrigerator. The implication is direct: the dominance of general trade, neighbourhood kiosks, open-air markets, and small-format souq retail, is not a temporary condition waiting to be replaced by modern trade. It is a structural reflection of where household infrastructure actually sits. In Kenya, higher refrigerator penetration has allowed supermarket chains like Carrefour and Naivas to build genuinely viable cold chain retail at scale in major cities. Ethiopia, at 1% national refrigerator penetration, is not yet at the threshold where a high-volume cold chain retail model is commercially sustainable beyond a handful of locations in central Addis Ababa. General trade will remain the dominant distribution channel for Ethiopian FMCG for the better part of the next decade. Any brand strategy, or media plan supporting it, that is built around modern trade assumptions is planning for a market that does not yet exist at the required scale.

Within this precision tier, digital behaviour is a useful additional layer. Mobile financial services penetration in urban Ethiopia has reached near-saturation of formal registration among working-age adults in major cities. Active recurring digital payment behaviour reliably identifies a household in the upper half of the urban distribution. As mobile money platforms deepen into lower income bands, that discriminating power will diffuse. The window in which digitally active financial behaviour cleanly identifies a commercially premium household is probably two to three years wide from this point. A digital behavioural overlay built on mobile app engagement and payment activity is a useful supplementary instrument for reaching this tier while that window is open. Supplementary is the right word. The reach base for this tier is still built on broadcast.

The formation tier

The fourth tier sits between the national awareness audience and the commercially addressable one. These are households scoring between 13 and 26, predominantly peri-urban and recently urbanising, aware of brands and aspiring to branded products, but without the physical infrastructure and income stability required to make those purchases a regular behaviour. Many are sachet consumers today. The conditions that would make them pantry consumers, reliable electricity, consistent formal income, improved housing, are forming around them as Ethiopia urbanises.

This segment represents approximately 19 to 21% of national households, roughly 5.6 to 6.2 million households. Advertising to this tier as a current pantry consumer audience is premature. Advertising to it as a brand awareness investment ahead of future market development has commercial logic, subject to two conditions: the brand needs an entry-level price point accessible to households approaching the Grade BC threshold, and the distribution needs to reach the secondary and tertiary urban areas where these households are settling. Without both conditions the investment is genuinely early. With both, the mental availability advantage among households first encountering branded categories can be significant. This is precisely the mechanism Byron Sharp’s work on penetration growth describes: brands grow by reaching buyers who are light or infrequent in the category, not by deepening spend among those already converted. The formation tier is Ethiopia’s largest pool of light future buyers, and most advertisers are not speaking to it at all.

If you are planning a campaign today and your brand has a three-year horizon, the Grade BC commercially addressable audience will be roughly 339,000 households larger by the time your next planning cycle closes. At a television-owning rate of 72 to 77%, that represents approximately 244,000 to 261,000 additional households entering the commercially addressable TV audience within that window, without any change in the urban electrification rate or income growth assumptions. Add the income growth trajectory and the grid expansion currently reaching secondary cities, and the true addition is higher.

The brand that has built recognition with these households during their transition reaches them with mental availability already established, before competitors have arrived. These households are arriving into urban life with smartphones before they arrive with televisions. The Socioeconomic Survey data from 2021/22, now three years old at the time of writing, almost certainly understates current smartphone penetration in the peri-urban band. Their first sustained exposure to branded advertising is often digital. The brand waiting for a television set to appear in these households before introducing itself has already missed the window.

Agency Takeaway: The formation tier is 5.6 to 6.2 million households that will be in your commercially addressable universe within two to three years. Most of them are currently being advertised to by nobody. The brand that reaches them first on mobile now, before they own a television, is building mental availability at zero competitive cost. That is not a charity investment. That is the cheapest market share available in Ethiopia today.

Curious Analyst Takeaway: The sachet-to-pantry transition is the most commercially significant household journey in the Ethiopian market right now. A household crosses from sachet consumer to pantry consumer not when its income rises past a certain number, but when its physical infrastructure, electricity, storage, housing stability, crosses a threshold that makes regular bulk purchasing rational. Electricity is usually the triggering variable. The grid expansion programme is therefore not just an infrastructure story. It is a consumer market creation story, and the brands positioned at the point of that transition will capture the share that forms there.

The digital-only trap

A reasonable response to all of this is: if the commercially addressable television audience is really only 4 to 5 million households, and digital reaches the same people with more precision and measurable response, why not put everything into digital and be done with it?

It is a coherent question. It leads somewhere that looks like efficiency and functions like a ceiling.

The first problem is that the commercially active digital audience in Ethiopia is not 32 million people. It is the same 4 to 5 million Grade BC and above households, now reached through a smaller screen with far more competition for their attention. A campaign that moves entirely to digital has not expanded its universe. It has changed the furniture in the same room. Impression counts on a dashboard can look substantial. The unique commercially active individuals behind those impressions are roughly the same population a television campaign would have reached, minus the passive viewing households who would have absorbed a brand name through repeated broadcast exposure without ever tapping a screen. Reach on a digital report and reach in the commercial sense of the word are not the same number.

The second problem is what happens to a brand message depending on the mental state of the person receiving it. Television is watched in a relaxed, passive state, typically in the evening, in a shared domestic setting, with sustained attention on a single screen. Influencer content and social feeds are consumed in a fragmented, active state, often while doing something else, with attention moving across dozens of stimuli in minutes. This distinction has measurable consequences for how well advertising is remembered.

A 2020 study published in the Journal of Advertising Research found that television advertising produced unaided brand recall rates of 40 to 50%, compared to 20 to 30% for digital formats across equivalent exposure frequencies. A Nielsen meta-analysis across 500 campaigns found that television generated 60% stronger long-term brand memory than digital video. The mechanism is not mysterious. Passive, emotionally engaged attention encodes brand information more deeply than divided, task-interrupted attention. A person watching television after dinner is neurologically more available to receive and retain a brand message than a person scrolling a feed between tasks. The message lands differently. It stays longer.

The third problem is that strong digital metrics can mask weak brand performance. Impressions, engagement rates, and follower counts are real measurements of real things. They are not measurements of mental availability, unaided recall at point of sale, or household-level commercial conversion. A campaign generating high engagement among a small, already-interested audience can look healthier than a broadcast campaign building slower recognition across a wider commercially addressable base, right up until the point where distribution reaches a new city and the television-familiar brand outsells the digitally optimised one because more people already know its name.

Agency Takeaway: Influencer recall at 72 hours is roughly half that of television at equivalent frequency. Your client’s digital engagement numbers are real. Their relationship to actual purchase conversion is weaker than the dashboard implies. A plan that uses television for memory and digital for reinforcement is not a compromise. It is the plan that actually works. The one that goes digital-only looks modern and performs below its potential.

Broadcaster Takeaway: The recall gap between television and digital is your most underused commercial argument. Television advertising produces unaided brand recall at 40 to 50%. Digital formats produce 20 to 30% at the same frequency. That difference is not a marginal advantage. It is the difference between a consumer who remembers your advertiser’s brand at the moment of purchase and one who liked a post and moved on. Lead with that number in every pitch.

Curious Analyst Takeaway: The reason television encodes memory more durably than digital is not about screen size or production quality. It is about the cognitive state of the viewer. Passive, relaxed attention, the kind a person has when watching television at home in the evening, is more receptive to memory formation than the fragmented, task-switching attention of someone scrolling a feed. Advertising research has been documenting this gap for two decades. The industry has largely ignored it in favour of what is easier to measure.

What the four tiers mean for a media plan

Nobody in the Ethiopian market is currently planning against this framework. The industry operates largely on the 17 million figure, applied uniformly, with digital added as a secondary line item when budgets allow. The analysis above is not a description of current practice. It is a description of what current practice is costing advertisers in unrealised commercial potential, and what it is costing broadcasters in credibility they have not yet been asked to defend.

The honest picture of Ethiopian media in 2026 is this. The national access audience sits somewhere between 12 and 17 million TV access points, with a parallel digital mass audience of 32 to 36 million mobile internet users that overlaps significantly at the top and extends broadly at the bottom. The commercially addressable Grade BC TV-owning urban audience is between 4.0 and 4.8 million households. The core FMCG precision audience is approximately 3.2 to 3.8 million households, the volume engine of category sales. And within that, the 294,000 to 330,000 refrigerator-owning households are the margin engine, the pantry consumers whose purchasing behaviour sustains premium pricing, bulk formats, and cold chain categories in a way no other segment currently can. The formation tier of 5.6 to 6.2 million peri-urban households represents the audience most advertisers are ignoring today and will be competing for in three years.

The question a media plan must answer before selecting a denominator is: what commercial behaviour am I trying to produce, and in which tier does that behaviour currently exist? For national brand awareness, category launches, and aspirational positioning ahead of distribution, the 12 to 17 million figure is a legitimate planning universe. For FMCG purchase conversion in the current cycle, the 4.0 to 4.8 million TV-owning Grade BC households is the correct denominator, with digital running alongside broadcast as a complementary layer rather than a substitute. For precision campaigns targeting pantry consumers across multiple branded categories, the 3.2 to 3.8 million core audience defines the target, narrowing further to the 294,000 to 330,000 refrigerator-owning households for cold chain and premium pack formats. For formation investment with a three-year payoff horizon, the emerging tier and its digital-first media behaviour defines the channel entry point, with television becoming the primary reach vehicle as those households cross the threshold.

The 17 million figure is a real measurement of a real thing. The analytical discipline is knowing precisely which thing it is measuring, and whether that thing is the right denominator for the objective on the brief.

Broadcaster Takeaway:

The 17 million figure is an asset, but only if it is explained correctly. Broadcasters who present their full satellite footprint as a single commercially active audience will eventually face the audit that forces a restatement. The more durable position is to sell the access audience for what it genuinely is, a national awareness reach vehicle with real value for the right objective, while simultaneously building the measurement that documents active urban viewership in the 4.0 to 4.8 million Grade BC tier. That second number is smaller, more defensible, and the one that keeps FMCG advertisers renewing year after year.

Nobody in the buying community is currently running this analysis. That is not a permanent condition. As marketing accountability increases and budget scrutiny tightens, the denominator question will eventually be asked. The broadcaster that has already answered it, with its own data, on its own terms, is in a fundamentally stronger position than the one reacting to a challenge it did not anticipate.

At 4.9% annual urban growth, Ethiopia is adding roughly 113,000 net new commercially addressable households every year, each one a pantry consumer in formation. A broadcaster that tracks this migration, by geography, by secondary city, by living standards score, can show advertisers a growing audience rather than a static one. The growth story is real. The data to tell it is available. The broadcaster that tells it owns the narrative.

Agency Takeaway:

The planning denominator you use determines the efficiency calculation you present to your client. A premium FMCG brand allocating budget against 17 million households is spreading its investment across a universe three to four times larger than the audience that is actually in a purchase cycle. That is not a reason to reduce spending. It is a reason to redirect it toward the tiers where the commercial return concentrates, and to make that reallocation the centrepiece of a smarter plan rather than a quieter budget.

Match the denominator to the objective. For brand awareness and distribution launch campaigns, the 12 to 17 million access audience serves. For FMCG purchase conversion, the correct denominator is the 4.0 to 4.8 million Grade BC TV-owning households, with digital running alongside as a complementary reinforcement layer, not a replacement for broadcast reach. For cold chain and premium categories, the 294,000 to 330,000 refrigerator-owning households is the planning anchor. These are the households stocking family-size formats at home, buying by brand across multiple categories, and sustaining the margin that makes premium products viable in this market. For formation investment, digital comes first in these households’ media lives and should come first in the channel plan, with television becoming the primary vehicle as they cross the Grade BC threshold.

The temptation to move budget entirely to digital because it produces measurable dashboards and lower apparent cost per impression is understandable. It is also, in most cases, the wrong move. The commercially active digital audience in this market is not a larger or cheaper version of the broadcast audience. It is largely the same people, in a noisier environment, with weaker memory encoding and lower unaided recall. Influencer content performs well on engagement and poorly on the durable brand memory that drives penetration growth. Television generates unaided recall rates roughly double those of digital formats at equivalent frequency, and that gap is what separates a brand people engage with online from a brand people reach for in a shop. The dashboard will look controlled. The brand building underneath a digital-only plan will be thinner, shallower, and slower than a plan that uses television for reach and memory, and digital to reinforce within the commercially active tier.

Penetration growth, which is how brands actually get bigger, requires reaching a broad commercially addressable population repeatedly until the brand occupies a place in memory that surfaces at the moment of purchase. In Ethiopia today, broadcast television is still the most efficient instrument for doing that at scale. Digital makes it sharper. It does not make it unnecessary.

To the challenger analysts: we exist in a data-thin market. The surveys are old, the sampling is imperfect, and some of the extrapolations here will feel forced to anyone who has spent time trying to build rigorous models on Ethiopian consumer data. That is a fair observation. It is not a reason to abandon the exercise. I approached this with a mathematical lens and a decade of media immersion, and the goal was never precision. It was direction.

STRATEX by Naz sits at the intersection of mathematics, AI, linguistics, and media, exploring the cognitive and behavioral foundations of how people, organizations, and markets work. If this kind of thinking is useful to you, subscribe below.