Netflix’s Afcon show: AI lessons for SA e-commerce

Netflix launching a daily Afcon highlights show (8am SAST, 22 Dec to 19 Jan) looks like a sports-media footnote. It isn’t. It’s a blueprint for how AI-powered personalization and content can pull people back into a product—without paying for the most expensive part of the market (live rights).

Most South African online businesses still treat content as “marketing stuff”: a few posts, a promo email, a campaign for Black Friday, then silence. Netflix is doing the opposite. It’s building a habit loop: yesterday’s action, packaged for this morning, recommended to the right viewer, in the right format.

This matters for our topic series—How AI Is Powering E-commerce and Digital Services in South Africa—because the same mechanics that make a highlights show sticky are the mechanics that make an online store profitable: relevance, timing, and repeat engagement.

Netflix isn’t chasing live rights—it's chasing daily habits

Netflix’s Afcon programme is intentionally “sports-adjacent”: highlights, analysis, interviews, fan reactions, atmosphere. It’s hosted by Robert Marawa and Minnie Dlamini, with analysis from Melissa Reddy, and it’s positioned as a catch-up companion rather than a live substitute.

That positioning is the point.

A lot of companies assume the only way to win is to buy the biggest, flashiest asset (in sports: live rights; in retail: the lowest price, the biggest range, the most discounts). Netflix is showing a more sustainable strategy: own the routine.

Here’s the underlying play:

• High frequency: daily episodes train repeat behaviour.
• Low friction: morning release time matches how fans plan their day.
• Broad accessibility: English audio plus English/French subtitles speaks to a pan-African audience.
• Data learning: each view teaches Netflix what audiences actually want.

For South African e-commerce and digital services, this maps directly to: use content to create repeat visits, then use AI to make those visits feel personal.

The real product is the recommendation engine (and it’s not just for streaming)

Netflix doesn’t win because it has “content”. It wins because it matches the right content to the right person at the right moment. The Afcon highlights show is another data-rich object in that system: who watches, when they stop, what they search next, what they rewatch, what they ignore.

South African retailers can apply the same idea with far less complexity than people assume.

What “personalization” should mean for SA online retail

Personalization isn’t “Hi Thabo” in an email. It’s changing what the customer sees and gets offered based on intent.

Practical examples that work well in SA e-commerce:

• Homepage re-ranking: reorder categories/products based on a shopper’s browsing and purchase history.
• Search that understands intent: if a customer types “Air fryer”, show accessories, recipe content, and delivery timelines—not just product tiles.
• Cart-aware content: if someone has soccer boots in cart, show a short “fit and sizing” guide and the return policy upfront.
• Repeat-buy reminders: for consumables (skin care, supplements, printer ink), predict replenishment windows and nudge at the right time.

Netflix’s Afcon show is “just highlights” on the surface. Underneath, it’s more training data for a system that increases watch time and retention. Retailers can do the same: every click can improve recommendations and reduce wasted ad spend.

If you’re paying for traffic but not learning from it, you’re buying the same customer twice.

“Highlights, not live” is a smart metaphor for e-commerce content strategy

Netflix avoided the cost and complexity of live sports rights while still capturing fan attention. South African e-commerce teams should copy this thinking: don’t start with the most expensive content formats.

A lot of brands jump straight to big-budget shoots and influencer campaigns. Then they can’t maintain consistency. Consistency is what builds the habit.

The e-commerce version of a daily highlights show

A daily highlights show has three ingredients: recap, perspective, and community. In retail, the equivalents are:

1. Recap: “What’s trending in your category right now?”
2. Perspective: “What should you buy and why?” (guides, comparisons, explainers)
3. Community: reviews, UGC, Q&A, returns experiences, staff picks

Concrete “daily highlights” formats that South African online stores can run sustainably:

• Daily deal with context: not just a price drop—explain why it’s good, who it’s for, and what to compare it against.
• Morning drop email: 3 items tailored to the customer’s category interest (not the whole catalogue).
• Short product explainers: 30–60 seconds, filmed simply, focused on one problem.
• Post-purchase “how to”: reduce returns and increase satisfaction (especially for electronics and appliances).

AI helps because it makes this content modular and targetable. You don’t need one perfect piece for everyone—you need many small pieces matched to the right customer.

AI use cases that mirror Netflix’s sports play (and drive revenue)

Netflix is experimenting around sports because it can measure behaviour quickly. That same measurement mindset is what makes AI useful in commerce.

1) AI-driven segmentation that updates in real time

Static segments (“women 25–34”) underperform because they don’t reflect intent. Better segments are behaviour-based:

• “Browsed running shoes twice this week”
• “Abandoned checkout after delivery fee shown”
• “Watches product videos before buying”

These segments should update automatically as customers behave.

2) Recommendations that factor in constraints (South Africa-specific)

South Africa adds constraints that recommendation engines must respect:

• Delivery coverage differences by area
• Load shedding disruption windows for call centres/warehouses
• Payment method preferences (cards, EFT, pay-in-store, BNPL)
• Price sensitivity and promo responsiveness

A useful recommendation system isn’t only “similar items”. It’s similar items that can actually be delivered quickly to that customer, at a price point they typically accept, with the payment method they use.

3) AI-generated content that stays on-brand (without becoming spam)

Generative AI is effective when it’s used to scale structure, not to flood the internet with generic words.

What I’ve found works in practice:

• Use AI to produce first drafts of product FAQs from manuals and specs.
• Use AI to create variant descriptions for different audiences (beginner vs enthusiast).
• Use AI to generate category landing copy that gets edited by a human for local relevance.

The rule: if the content doesn’t reduce customer effort (fewer questions, faster decisions, fewer returns), don’t publish it.

4) Customer service automation that feels like a good concierge

Netflix doesn’t call you to ask what you want; it quietly puts the right thing in front of you. That’s the vibe SA digital services should aim for.

Good AI support in e-commerce includes:

• Order tracking bots that handle the top 20% of queries that create 80% of ticket volume
• Returns flow assistants that explain eligibility clearly (and reduce agent back-and-forth)
• Product compatibility checkers (for chargers, ink, laptop RAM, phone cases)

This is where leads get won: fast, clear service turns hesitant shoppers into repeat customers.

Infrastructure is the quiet enabler: streaming and shopping run on the same rails

Netflix can publish an 8am SAST show reliably because the underlying digital stack is strong: content delivery, data pipelines, analytics, and increasingly, advertising technology.

South African e-commerce and digital services rely on the same rails:

• Cloud platforms and CDNs for speed
• Payment gateways and fraud controls
• Data warehouses and event tracking
• Marketing automation and attribution

If your site is slow, your catalogue data is messy, and your tracking is incomplete, AI won’t save you. It will amplify the chaos.

A practical 2026 readiness checklist (useful right after the festive rush):

1. Fix product data: clean titles, attributes, images, stock accuracy.
2. Track events properly: view, search, add-to-cart, checkout, purchase, returns.
3. Create 20–30 modular content assets per category (guides, FAQs, short videos).
4. Personalize one surface first (homepage, search, or email)—then expand.
5. Measure retention, not just conversion: repeat rate over 30/60/90 days.

A “People Also Ask” set your team should answer internally

Does AI personalization only work for big retailers? No. Smaller stores often win faster because they can implement simpler rules and iterate quickly. Start with email recommendations and on-site search improvements.

What’s the fastest AI win in South African e-commerce? Cart and checkout optimization using behavioural data: reduce delivery-fee shock, clarify returns, and recommend alternatives when an item is out of stock.

How do you avoid creepy personalization? Personalize based on shopping behaviour on your site/app, not sensitive personal traits. Keep controls visible: preference centres, easy opt-outs, clear explanations.

How do you know content is working? If content increases either (1) conversion rate for engaged viewers, or (2) repeat visit rate, it’s working. If it only increases impressions, it’s noise.

What Netflix’s Afcon move should trigger for SA businesses

Netflix didn’t need live Afcon streaming to create value. It needed a format that builds routine, creates emotion, and feeds the recommendation engine.

South African e-commerce and digital service providers should take the hint: content and AI aren’t separate projects. Content creates engagement; AI decides what to show, to whom, and when. Together, they reduce acquisition costs and increase customer lifetime value.

If you’re planning your 2026 pipeline, I’d make one bet: habit beats hype. Daily (or weekly) repeatable content, matched with practical AI personalization, will outperform another once-off campaign.

Where could your business create a “highlights show” experience—something customers actually look for each morning—inside your store or app?