Paid Social and M-Pesa Checkout Optimisation for a Nairobi Fashion Ecommerce Brand

Context

Our client is a direct to consumer fashion brand based in Nairobi, selling women's and men's ready to wear pieces through its own Shopify style storefront and an active Instagram and Facebook presence. The brand had built a loyal following through influencer seeding and studio shoots around Kilimani and Westlands, but sales were inconsistent. Most months relied on flash sale spikes rather than a repeatable acquisition engine. The founder came to us with a clear brief: turn social media attention into paid, measurable revenue, and stop losing customers at the point of payment.

Like most Kenyan ecommerce operators, the brand's checkout flow leaned almost entirely on M-Pesa, with a smaller share of orders paid on delivery in Nairobi and a card option that barely got used. Their monthly paid media budget was modest, in the range of KES 150,000 to KES 250,000, spread thinly across boosted posts with no real campaign structure.

The Problem

Three issues compounded each other. First, the paid social account had no funnel logic. Every campaign pushed cold traffic straight to product pages with a single generic "shop now" message, regardless of whether the audience had ever engaged with the brand. Second, checkout abandonment was severe. Our early analytics review showed a cart to purchase completion rate under 18 percent, with the steepest drop happening exactly at the M-Pesa STK push step. Customers were opening the prompt on their phones, getting distracted or confused by the payment amount format, and simply not completing it within the short window before the prompt expired. Third, there was no retention layer. Every sale was treated as a one off, with no structured way to bring buyers back for a second purchase, despite fashion being a naturally repeat category.

We also found that a meaningful share of traffic was arriving on low end Android devices over Safaricom mobile data, meaning page weight and load speed were quietly taxing every campaign's return before checkout even began.

What We Did

1. Audited the full funnel from ad click to M-Pesa confirmation SMS. We mapped every step, timed load speeds on 3G and 4G Safaricom connections, and recorded session replays to see exactly where shoppers hesitated or exited during the STK push flow.

2. Rebuilt the paid social account structure around funnel stages. We split campaigns into cold awareness content built from short form video of the studio shoots, warm retargeting for people who had viewed products or added to cart, and a dedicated cart abandonment sequence timed to fire within two hours of drop off, when purchase intent from the original ad was still fresh.

3. Simplified and clarified the M-Pesa checkout experience. We shortened the checkout form, displayed the exact KES amount and till number before the STK prompt fired so there were no surprises, and added a visible countdown with a clear "resend prompt" button so customers who missed the first push did not have to restart the entire order.

4. Introduced order confirmation and delivery expectation messaging via SMS and WhatsApp. For Nairobi orders this included realistic delivery windows based on which part of the city the customer was in, since delivery to areas like Ruaka or Rongai understandably takes longer than delivery within the CBD or Kilimani.

5. Built a lightweight loyalty and win back layer. We tagged every customer's first purchase category and sent a retargeting sequence roughly three to four weeks later featuring complementary pieces, along with a small loyalty incentive for a second order.

6. Set up proper attribution and reporting. We connected ad spend, M-Pesa transaction data, and delivery status into a single weekly dashboard so the founder could see cost per acquisition and repeat purchase rate at a glance instead of relying on Meta's in platform numbers alone.

Results

Over four months, blended return on ad spend rose from roughly 1.6x to 3.9x. Cost per acquisition fell by 41 percent as the funnel restructuring meant cold budget was no longer competing with retargeting budget for the same conversion event. The checkout simplification had the single biggest measurable impact on revenue: STK push completion rate rose from under 18 percent to just over 52 percent, meaning more than double the customers who reached checkout actually paid. Repeat purchase rate within 60 days climbed 2.6x, turning what had been an almost entirely first time buyer business into one with a genuine, if still early, retention engine. Average order value also rose modestly as the win back sequences favoured slightly higher priced complementary items rather than discounted basics.

What We Would Do Differently

We would have prioritised the checkout audit even earlier. We spent the first three weeks focused on paid social structure before realising the checkout leak was masking a large share of that work's impact, and fixing it first would have let us measure true funnel improvements from campaign changes without a moving baseline underneath them. We would also push harder, earlier, for a proper card and bank transfer fallback alongside M-Pesa. A small but consistent share of higher value orders came from customers who preferred not to use mobile money for larger transactions, and we only added that option in month three. Finally, we underestimated how much delivery transparency mattered to repeat purchase rate. Once we started segmenting Nairobi delivery estimates by area, we saw fewer support queries and better second purchase behaviour, and in hindsight this should have been part of the checkout rebuild from day one rather than a month three addition.