Declining top-up volume is one of those problems that looks straightforward on a dashboard. Volume is down. Share of wallet is moving to the competitor. The working hypothesis is usually a product problem or a pricing problem, and the response is usually a campaign or a promotion. In one market, that instinct would have been wrong. The volume decline was real, and the share loss was real, but neither was caused by what the product or marketing teams could see from where they were sitting. The problem was structural, sitting at the bottom of the distribution chain, at the exact point where a customer walks up to an agent and asks to top up. Understanding why that matters requires understanding how prepaid top-up distribution actually works, and where it can quietly break.

The distribution chain and how it creates value

In markets where cash-based, in-person top-up remains the dominant recharge channel - which describes a significant portion of the prepaid base across many Caribbean and LatAm markets - distribution is not a support function. It is the product. A customer who cannot find a convenient top-up point within a reasonable distance of where they live or work does not go out of their way to find one. They either churn to a competitor with better local coverage or they reduce their recharge frequency. Both outcomes are visible in the volume data. Neither is visible as a distribution problem until you look specifically for it. The chain in this type of market typically runs through several tiers. A small number of primary distributors hold the main commercial relationship with the operator, receiving a discount rate on airtime they move through their networks. Below them sit sub-distributors, who take a portion of that margin and supply the merchants and agents at street level. Those merchants - the person at the counter of a pharmacy, a supermarket, a small shop - are the ones actually facing the customer. They are also, in most cases, stocking and selling the competitor's top-up product alongside the operator's. When a customer walks in, the merchant's recommendation or default behaviour will follow the margin. Whichever product pays them better at the point of sale is the one they will quietly favour. This creates a flywheel that can run in either direction. When it runs forward: more agents in more locations generate more convenience, more convenience drives more top-up volume, more volume generates more commission income through the chain, more income makes the agent role more attractive, and more agents join or stay active. The loop reinforces itself. When it runs backward, the dynamic is the same but inverted. Falling margin at the end-merchant level means agents favour the competitor. Fewer top-ups get sold. Less commission flows back through the chain. The next agent has even less incentive to prioritise the product. Coverage quietly degrades - not because agents have left but because the ones who remain are choosing the competitor at the moment of truth. The critical insight is that this can happen while the headline distributor discount rate looks perfectly adequate. The operator sees a reasonable margin going to the top of the chain and assumes it is working. What they cannot see without looking specifically for it is whether that margin is actually reaching the merchant at the bottom.

The diagnostic

The diagnostic was built around three data layers, overlaid at cell-site level across the market. Top-up transaction volume by cell site gave a picture of where revenue was actually being generated and where it was falling. Agent and point-of-sale concentration in each geography showed how many active top-up points existed relative to the population and device density in that area. Connected device data - meaning how many active prepaid SIMs were in each geography - gave a demand baseline to compare against the supply of coverage. The combination revealed something a national average would never show. Coverage was not uniformly degrading. Specific geographies, largely outside the main urban corridors, in areas with lower smartphone penetration and less access to digital recharge channels, had seen the number of active top-up points decline sharply over time. In these areas, in-person cash top-up was not a secondary option. It was the only practical channel available to a meaningful share of the prepaid base. The coverage gap was not an inconvenience. It was a complete removal of the ability to recharge. The competitive read added another dimension. The working assumption inside the team - based on volume and share-of-wallet trend data - was that the competitor's local coverage lagged the operator's overall but was closing the gap. This assumption was inferred from internal data rather than independently verified against competitor coverage information, and should be understood as a hypothesis that shaped urgency rather than a confirmed fact. What it did was narrow the window: if the assumption held, the moment to act was now, not after parity had been reached. The sub-distribution chain was then mapped to understand what commission was actually reaching the end-merchant, rather than what the primary distributor contract suggested should be flowing. The result confirmed the hypothesis and put a number to the leak. Sub-distributors were earning an average discount of 8.5% on the airtime they moved. Of that, only 1 to 2 percentage points was reaching the agents, mom-and-pop shops, and point-of-sale operators who sold directly to the customer. Nearly all of the margin was being absorbed one tier above the person actually deciding which product to hand across the counter. The top-line rate looked adequate on paper. The end-merchant had almost no financial reason to push this operator's product over the alternative.

The fix

The diagnostic had identified precisely where the chain was breaking, so the interventions were sequenced to match. The commission restructuring came first, since the margin leak at the end-merchant level was the confirmed root cause of the volume decline. The own-agent build-out ran in parallel in the gap geographies identified by the cell-site mapping, while hiring and training were completed. The organisational redesign and daily cadence sat underneath both, giving the team the visibility to monitor recovery in real time as each piece landed. The flat average of 8.5% at the sub-distributor level was replaced with a tiered, performance-linked structure: 7.5% at 90% of target, 8.5% at 100%, 9.0% at 110%, and 9.5% at 120% and above. The same tiered logic was extended to the operator's own agents and merchants, who moved from a baseline of 1 to 2% up to the same 8.5 to 9.5% band at and above target. The intent was to make every layer of the chain - not just the top - a direct stakeholder in hitting volume targets. A sub-distributor sitting below 90% now had a visible, immediate reason to push their own downstream network harder rather than coast. The pressure was designed to flow downward through the chain automatically, rather than requiring the operator to manage thousands of individual merchant relationships directly. It is worth being clear about what this fix costs, not just what it fixes. The new structure pays out more per unit of top-up sold than the prior flat rate, both at the top-performing tier and at the low-performing end. The incremental commission cost, and the separate hiring and platform development costs of the own-agent build-out, need to be netted against the revenue and share recovered to produce a complete picture of return. The fix addresses the right structural problem. The cost-adjusted return is the next thing to establish. The own-agent build-out was targeted specifically at the geographies where coverage had collapsed and where no existing distributor network was actively serving customers, rather than layering new agents on top of areas already covered. The intent behind this placement decision was to drive healthy competition rather than cannibalization: agents placed in genuinely unserved territory would, in principle, create pressure on existing distributors to extend into those areas themselves rather than simply displacing their existing volume. The placement logic follows directly from the gap map the diagnostic produced. The third intervention addressed a structural visibility problem that the commission and coverage fixes alone could not solve. Oversight of the sales and distribution function had sat with a single manager responsible for the entire territory, with no practical way to maintain daily contact with the ground level across a dispersed geography. Issues at the merchant and agent level - precisely where the root cause of the volume decline was eventually found - could go unnoticed for weeks between scheduled visits. The sales organisation was restructured so each geographic area had its own lead, accountable for every channel operating in that territory: stores, distribution partners, full service dealers, and the operator's own agents. Each lead was supported by a dedicated operations agent and sales agent, giving every area its own small execution team. A daily morning call was introduced to review each area and sub-area against the previous day's target, identifying immediately where performance had been met or missed and what needed to be adjusted - whether a below-the-line promotion, a CVM targeting change, or a marketing calibration for that specific geography. A problem that previously could compound undetected for weeks now surfaced within twenty-four hours and could be corrected the same day.

The results

Because the fixes were sequenced from a specific diagnostic rather than deployed as a broad simultaneous program, the results that followed are consistent with the causal chain the data identified. Top-up volume recovered as coverage gaps closed and end-merchant commission economics became competitive. At the level of the broader program, subscription services including postpaid and FTTH reached 120% of target month over month. Revenue share grew 8% within six months. Market share grew 1% within eight months. The subscription, revenue, and market-share figures reflect the full program including the sales organisation redesign, which also strengthened postpaid and FTTH channel performance. The top-up recovery and the broader commercial results moved together because the same diagnostic discipline, the same sequenced intervention logic, and the same daily visibility structure governed all of it.

What this diagnostic transfers to

The mechanics of this problem are not specific to prepaid telecoms. Any business that moves its product to end customers through a multi-tier distribution chain - and where those end distributors are simultaneously selling a competitor's product - is exposed to the same flywheel dynamic. The specific questions to ask are consistent across sectors. Is the incentive at the top of the chain actually reaching the person making the recommendation at the bottom, or is it being absorbed in the middle? Where has coverage quietly degraded below a threshold that matters to the customer, in ways a national average will never surface? And does the organisation have the visibility structure to detect a local problem before it becomes a national one? These are questions that apply equally to FMCG distribution networks, mobile money agent infrastructure, pharmaceutical wholesale chains, and a range of other channel-dependent businesses. The diagnostic tools differ by context. The underlying commercial logic is the same. A distribution flywheel that has started running backward will not be fixed by a marketing campaign upstream. It will be fixed by understanding where in the chain the reinforcing loop broke, and intervening at that specific point. Getting the incentives right matters. Getting the visibility right matters just as much. National averages hide this almost every time. The problem is always local.