The technology behind Takealot's Personal Shopper programme

Some of the most consequential e-commerce in South Africa doesn't happen on a website at all - it happens through a person. Takealot's Personal Shopper programme connects communities across the country to online shopping through local shoppers, who place orders on behalf of customers and earn commission for it, and since its public launch in May 2024 it has become a visible part of Takealot Group's strategy for reaching the many South Africans who have never transacted online. A neighbour who knows the platform becomes the bridge to it, and that simple idea has grown into a network of more than 14,000 shoppers, and growing.

What the public didn't see was the machinery behind that growth - because for its first stretch, there barely was any. The programme ran on spreadsheets, informal channels and the sheer diligence of the people operating it, and while the operations worked, the infrastructure could not keep up with what the network was becoming. Takealot needed a proper technology layer: built to enterprise-grade security standards, integrated into its existing systems, and able to serve three very different user groups from day one.

A first for Takealot

The decision that followed was a notable one for a company that builds its own technology as a matter of identity: the Personal Shopper platform became Takealot's first outsourced product build, and Touchfoundry became the partner trusted with it. The brief had no spec attached - strategy, product development and build would need to run as a single stream - and the commercial shape mattered as much as the technical one: the work was costed to scope and delivered on a fixed budget and timeline, with the delivery risk carried on our side. Takealot knew what they were getting, what it would cost and how long it would take before a line of code was written.

Rather than working at arm's length, we embedded inside Takealot's product function - inside their architecture and governance, alongside their product owners, technical leads, security and compliance teams - and delivered as one full team spanning strategy, experience, technology and data, with the delivery cadence running to our rhythm rather than being split across two. For an enterprise, that's the real purchase: not an app, but a proven team that plugs into your standards without six months of recruiting or a procession of coordinated suppliers.

Two products, three user groups

The platform's shape follows the programme's shape. Three user groups needed serving; two of them needed a product surface, and the third - the Takealot business itself - is served through direct integration into Takealot's existing systems rather than another interface to staff and maintain. That restraint is a design decision we'd defend anywhere: the best product for a business layer is often no product at all.

For the shoppers in the field, there's the Personal Shopper app - an installable PWA behind Takealot single sign-on that gives each shopper a single view of their business. Commission tracking and sales goals, tier progression with a live leaderboard, a built-in CRM for their clients and orders, and deals that surface in real time with native sharing - with orders syncing from Takealot hourly, so the app and the platform never disagree about the numbers that matter to someone's income.

For the team running the network, there's the Community Manager dashboard - the control room. Shopper onboarding (CSV upload included, because real operations arrive in spreadsheets), cohort assignment and tier configuration, scheduled deals and broadcasts targeted to tiers or individuals, support handled through integrated Zendesk ticketing, and event-level Google Analytics underneath it all. Both products share one authentication, one data infrastructure and one design system, which is the unglamorous foundation that makes the whole thing feel like a platform rather than two apps in a trench coat.

The best product for the third user group was no product at all - just integration, done properly.

Six months, on time and on budget

The delivery ran from a standing start to live in six months, on the fixed budget, with strategy and build overlapping as one stream rather than queuing as phases. From pilot - the platform in the hands of its first shoppers - the work moved straight into full operations without a handover, because the team that built it is the team that runs it: critical fixes, feature iterations and improvements tracked through one channel, with the product context carried intact from the first workshop into live service.

“We are glad to have found a partner that understands the market; contextualises technology to South African problems; talks the language that we talk and moves at the speed at which we move. That has been the win-win recipe for us.”

Where it stands

Today the platform supports more than 40,000 transactions a month across the network's 14,000+ shoppers, and it keeps growing under the same managed-service arrangement - build into run, one partner from strategy through to support. For the programme, the technology layer that once lived in spreadsheets is now infrastructure; for the shoppers, the platform is the single place their business lives; and for us, it stands as the fullest expression yet of the way we prefer to work: embedded, accountable to a fixed promise, and still in the room long after launch.

There's a broader note worth ending on. Enterprise South Africa has historically treated product outsourcing as a risk to be managed rather than a capability to be bought, and engagements like this one make the counter-argument in production: with the right model (fixed cost, carried risk, embedded delivery, build into run), an outside team can be held to inside standards - and move faster precisely because it arrives whole. We suspect Takealot won't be the last to test that proposition, and the full case study below goes deeper on how the model works.