A Kenyan company increased per-customer sales by 15%, extended daily cooking durations, boosted active usage days, and shifted a portion of customers from charcoal to liquefied petroleum gas (LPG) using the Advanced Jobs to Be Done (AJTBD) methodology.

Company: M-Gas—a provider of clean cooking fuel for Kenyan households—operates on a pay-as-you-cook model with smart gas meters. It competes with traditional fuels (charcoal, firewood, kerosene) and modern alternatives like bioethanol (KOKO).

This case was shared by Den, Head of Product Development at M-Gas. Den narrates, with commentary from Ivan.

Company and Market Context: Clean Cooking in Kenya's Micro-Economy

As Head of Product Development, Den oversees product evolution. Based in London (the global headquarters), Den travels to Kenya for direct customer and team interactions.

Kenya, with over 50 million people and moderate income levels, sees many households—especially in rural areas—relying on firewood or charcoal for cooking: access to clean fuels was only about 24% in 2021. Families earning $100–300 per month plan meals from tight daily budgets, avoiding large upfront costs (a 6 kg gas canister costs $10–12; a full cylinder $18–20—for context, rent runs $35–50 monthly).

Mission

Deliver clean, incrementally paid cooking fuel to make LPG a viable alternative to charcoal and firewood. This "micro-dosing" approach aligns with Kenya's kadogo economy—akin to pay-per-use models in Western markets, like top-up mobile plans or micro-transactions in apps.

Business Model

M-Gas operates a pay-as-you-cook model: A stove and smart-metered cylinder, developed in-house, are installed in customers’ homes. Users top up small amounts via M-Pesa (Kenya’s mobile money service), and the meter releases exactly the paid-for gas—for instance, 10–15 Kenyan shillings (~$0.10) covers breakfast prep. Indirect competitors include charcoal, firewood, kerosene, and KOKO (bioethanol, refilled at kiosks).

Personal Insight

Den shares: With a background in strategy and corporate finance but no prior product experience, I faced real "imposter syndrome." Proving my ability to stakeholders and myself was critical. The AJTBD methodology’s clarity and practicality became a game-changer.

The Challenge: Drive Sales Growth Without Raising Churn

The goal was to boost per-customer sales and revenue without increasing churn. In Kenya’s context, target households face irregular, low incomes and inconsistent consumption patterns.

A key issue: About 50% of customers blend M-Gas with other fuels (charcoal, firewood, kerosene, or spare LPG cylinders). Heavy investments in meters and operations mean low usage leads to losses for some clients. Capturing more "wallet share" and positioning LPG as a true charcoal rival were critical.

Initial Hurdles: Customers Couldn’t See the Value

Previously, M-Gas relied on discounts and ad-hoc pricing tests without a dedicated product development function.

Customers struggled to compare M-Gas costs with rivals. A bucket of charcoal has a clear price, while M-Gas expenses varied by session length, flame intensity, stove wear, and even altitude. Budgeting was reactive—tallying spend until a 13-kg cylinder swap or estimating per dish ("tea costs roughly this; that stew, about that").

Den notes: Customer behaviors varied widely, with gas consumption differing by factors of two or more. Data was "noisy" (inconsistently logged by field teams) and showed no clear consumption patterns. A behavioral pricing test deepened concerns—stakeholders pushed for tiered pricing, but customer responses to changes barely impacted sales or revenue.

Despite M-Gas being cheaper than charcoal in most regions and faster, distrust in unpredictable gas costs led customers to default to charcoal.

Breakthrough via In-Person Interviews

In early 2024, Den began conducting face-to-face interviews in customers’ homes—modest apartments renting for $30–40/month in areas rarely visited by outsiders. Many clients spoke no English, so Den learned basic Swahili on the fly (e.g., kupika haraka—the job of "cooking quickly").

Qualitative interviews uncovered 23 jobs, validated quantitatively on a 1,000+ sample. The team targeted a frequent, high-priority job: Cook with whatever money I have right now ("no matter how little"), plus jobs tied to extended cooking sessions.

Den’s hypothesis: Focus on jobs where charcoal "wins" (perceived as cheaper), informed directly by early interviews.

Solution: M-Gas Minutes

M-Gas launched M-Gas Minutes—a 70-minute cooking bundle for 49 KES (~$0.40). Minutes are valid for 24 hours, splittable across meals (breakfast, lunch, dinner), and stackable. It’s an add-on to the gram-based model.

Customer Value

  • Predictable Pricing: A flat daily cost eliminates mental math on per-kilo expenses or stress over unpredictable bills, regardless of the dish.
  • Charcoal Comparison: A charcoal bucket costs 80–100 KES; even two Minute packs match or beat it while being faster and cleaner.
  • Accessibility: Capped under 50 KES, it aligns with the core job of affordable fuel.

For slow-cook dishes like githeri (corn-and-bean stew), families often chose charcoal pre-Minutes. The pilot tested 60- and 90-minute options; analytics and cooking studies set ~70 minutes as the "daily standard" for full meals. The marketing message: Switch all cooking to M-Gas Minutes—it covers your whole day.

Ivan’s Commentary

Under the gram-based model, customers struggled to predict cooking costs, adding extra jobs (e.g., mid-cook expense tracking). M-Gas Minutes eliminates low-level work with fixed, predictable pricing, simplifying the core job of cooking with available funds.

Implementation Challenges

Adopting AJTBD was straightforward with a lean team, stakeholder support, and research team enthusiasm after pilot interviews.

However, M-Gas Minutes faced hurdles:

  • Technical: New meter firmware required months of development, even for testing.
  • Awareness: Operational noise meant many customers didn’t know about or how to activate Minutes. Den’s Kenya trip revealed adopters’ enthusiasm and non-triers’ eagerness.
  • Distribution: SMS and field teams (during cylinder/meter swaps) were key channels, but staff were stretched thin. Aggressive follow-through—metrics, feedback, and centralized calls—drove adoption.

Due to tech constraints, firmware covers only ~25% of users, limiting broad marketing.

Results

Metrics:

    • Uplift: At least +15% per-customer sales post-spring rollout (vs. matched control group).
    • Behavioral Shifts: Increased daily cooking duration and active days, signaling a shift from charcoal and alternatives.

Den’s Reflection: “Launching M-Gas Minutes built confidence and laid the groundwork for future innovations.”