Plus: Meta’s Twitter competitor, how you trained AI & building things that don’t scale (on purpose).

Newsletter

July 6, 2023

Hi there,

Think you’ve mastered AI prompt injection? (Tricking an AI into revealing things they shouldn’t.) AI safety company Lakera developed the Gandalf game, where you try to trick the AI into revealing its secret through multiple levels. Go get your AI hack on.

In this Open Letter:

Big ones: Saving 15M’s daily commute.
SA crypto exchange crackdown, Meta’s Twitter competitor & thanks for training our AI.
Out of the box: How to build a network-effect product.
Get practical: Building things that don’t scale with Ben Blaine.

TRENDING NOW

15 Million People are About to Lose their Only Means of Daily Transport

And there’s much to be done to stop it…

In the sphere of urban planning, it's become increasingly important to develop mixed-class neighbourhoods. For two reasons: combating classism and reducing the burden on our transport systems.

Living close to work means less commuting, promoting a higher quality of life. So proper city planning not only aims to reduce dependence on transport infrastructure but also to create healthier, more connected communities.

*Smart cities? No roads, just like Wakanda.*

Remember apartheid's approach, which pushed people out of cities and into townships? It essentially undermined these principles of community development. In its wake, the minibus taxi industry emerged, serving the transport needs of the populace (circa 15 million people per day) and creating profitable enterprises in the process.

The bump in the road

During our recent chat with LÜLA CEO Velani Mboweni on How Would You Build It, we learnt that some 60% of South Africans depend on minibus taxis for their primary transport. But, unlike bus services, taxis don’t receive government subsidies. And, like it or not, our economy is deeply reliant on the minibus taxi industry for daily transport.

Now, there's a growing concern about taxi owners who are struggling. And minibus financing specialist SA Taxi is also feeling the pressure. Its parent company, Transaction Capital (who also owns WeBuyCars)’s share price is down substantially and SA Taxi has undergone major restructuring to keep it going.

Consider that SA Taxi finances over 36’000 of the estimated 250’000 taxis in South Africa (about 15%, or every 7th taxi you see on the road). And if this company is struggling, it is a reflection of broader issues.

Cause and effect

The crisis is not due to a lack of industry-specific innovation. We were impressed when we visited SA Taxi’s operations a few years ago – their understanding of the taxi business and ability to offer great value whilst reducing risks and costs way surpasses anything traditional banks could do.

They finance your taxi for specific routes
- Knowing how many commuters and taxis are on every route, they know whether you’ll make money (and could thus afford the loan).
- Drive a different route and they cut you off remotely with a smart GPS device.
They know one day off the road is one step closer to defaulting on the loan, so have all kinds of initiatives to help get taxis back on the road safely in record time
- They offer custom insurance that processes claims quicker.
- They train township mechanics to fix certain things on taxis, improving the quality of work (also uplifting them in the process).
Repossessions have less impact on the business, as they would refurbish and refinance the taxi to a new owner – something a traditional bank can’t do well.

Despite these innovative approaches, several factors have contributed to the taxi industry battling to stay afloat:

The cost of vehicle ownership (repayments, insurance, maintenance and fuel) increased a staggering 48% since 2019.
Flooding at Toyota’s manufacturing site has created additional challenges.
Other headwinds include rising interest rates and fuel costs.
Lower commuter volumes due to slow economic growth and remote working policies.
Fleet owners are unable to increase fares for cash-strapped commuters.
Loadshedding is slowing traffic, reducing the number of trips drivers can make decreasing income.
Less qualifying buyers for repossessed and refurbished vehicles.

Light at the end of the tunnel

While the situation is concerning, the taxi industry has shown resilience in the past. There's still room for optimisation within the current model, and technology could play a significant role.

Some of the obvious challenges include:

It normally takes two people to operate a taxi – one to drive and one to collect funds and check admission. Salary is by far the biggest cost of operating transport, so being able to remove one of these will go a long way in saving costs.
Taxis often wait till they reach a certain amount of passengers before setting off – which sounds logical to maximise income, but it's not that simple (you’re making current passengers wait).
Although they’ve taken steps to try to formalise the industry, better coordination between different operators could improve efficiency and margins for everyone.

*Might have to call in sick if this goes on much longer.*

Opportunity time

Looking forward, there's a raft of opportunities for the taxi industry that could streamline operations and boost viability:

Tax compliance is key for taxis to start receiving subsidies. In 2021, it was reported that this R90bn industry paid only R5m in company tax. Although some in the industry might already be paying employment or personal income tax, there's a need to formalise this process further. Subsidies could incentivize tax compliance, with technology playing a pivotal role in this transformation.
Enhanced route planning is a potential game-changer. The combination of cameras, IoT, and AI can facilitate improved route planning and scheduling, optimizing earnings per trip. Cape Town startup, Quickloc8, is already venturing into this territory.
Loop Mobility is endeavouring to increase demand by refining the user experience. They're offering an "Uber-like" service for minibuses, handling payments and offering insurance.
There's an opportunity to introduce subscription model rides for a consortium of minibuses, which could stabilize cash flow, making it more predictable.
Even though cash transactions still dominate the minibus taxi world, improved payment management could mitigate default risks and foster better planning. This is particularly relevant when the taxi owner is not the driver. By leveraging technology and digital payments, revenue from trips can be secured, potentially allowing operations to be handled solely by the driver.

For years, tech adoption had been lacklustre in this space, perhaps now with sufficient pressure, the time to use technology to optimise is finally here. We are excited to see this industry go from strength to strength.

Need an inside look into the tax industry? Hit reply, we might be able to organise it for you…

IN SHORT

🥵 Hot stuff. Records are meant to be broken (or are they)? Monday was the hottest recorded day globally with an average global temperature of 17 degrees Celsius. Probs something to do with that pesky El Niño thing we spoke about in a previous Open Letter.

🧵 Stitched up. Meta’s Twitter competitor (called Threads) is set to release today. According to a source inside Meta, they believe a version of Twitter run by a “sane person” is what the world needs. Depending on your definition of sane, it's either very good or very bad. Got an iPhone? The app will appear here.

🥊 Beat down. DStv’s owner MultiChoice’s share price took a beating after JP Morgan Chase & Co downgraded its rating of the company. This while French premium TV channel Canal+ increased their stake to 31.7% — one step closer to being forced to offer a buyout to other shareholders (which is at 35%).

🤑 Got crypto? SA exchanges might soon be forced to shut down if they don’t have licenses. SA’s Financial Sector Conduct Authority (FSCA) is pushing for regulation and SA is set to become the first African country to require digital asset exchanges to get licenses.

🎉 AI Trainer. Congrats on officially becoming an AI Trainer. Well, your publicly available information anyway. Google just updated its privacy policy saying they use publicly available info to help train their AI models. Time to go and update our LinkedIn Bios.

BUILDER’S CORNER

How to Build a Network-Effect Product

Want to build the next TikTok, Facebook, Uber or Airbnb? You know, products whose true value only unlocks once you’ve onboarded a certain number of users/stakeholders.

Well, the trick to building a network effect product is to NOT DO IT (at first, at least).

Method in the Madness

According to Y Combinator co-founder Paul Graham, and his 2013 essay that’s become startup folklore, it’s much better to build things that don’t scale at first.

What that means is: Network-effect products are complex, costly and super risky to build. If you don’t nail it first-time, you waste years of time and resources and so much money your investors won’t ever want to touch you again – that’s why most won’t invest in network-effect products anymore.

It’s best to build a non-scaling version first.

For example: Instead of trying to build a fancy new product

see if you can deliver the service using simple tech like a spreadsheet first
or see if you can build the community it will eventually serve, first.

Then, only once you have the community and service running so smoothly you can’t manage it anymore, that’s when you consider building the product. (It’s more common than you think: EskomSePush also started as spreadsheets first.)

So, how do you build for not scaling? Well, we were lucky enough to have an expert at that in our “How Would You Build It” podcast this week – check it out down below.

For now…

5 Steps for Building Solutions That Don’t Scale

Start with a deliberately narrow market
Instead of targeting a broad audience, focus on a specific subset of users who are likely to adopt your product or service quickly. By concentrating your efforts on a niche market, you can create a critical mass of users and gain traction more rapidly. Expand to a broader market once you have a solid foundation.
Recruit early adopters manually
Engage in personal outreach, offer hands-on setup assistance, and take advantage of your existing network and connections to acquire users.
Focus on user delight
Go above and beyond to make your early users happy. Provide exceptional customer service, personalize the user experience, and find creative ways to exceed their expectations. Aim to make signing up with your startup one of the best choices they've ever made.
Obsess over delighting users
Strive to create an insanely great experience for your users, even in the early stages when your product may be incomplete or buggy. Hint: it’s not your tech. Ask yourself – what is a 5-star experience? Pay attention to the details, iterate based on user feedback, and be attentive to their needs. Continuously seek ways to improve and make the user experience exceptional.
Engage as consultants for a single user
Treat a single user as a consulting client and build a tailored solution specifically for their needs. Continuously iterate and refine your product until it perfectly meets their requirements. By solving their problem effectively, you increase the chances of attracting similar users and expanding into adjacent territories.

Want to delve deeper into the topic? Check out our podcast below – it’s seriously great for when you want to build smart and well.

Stuck on how to “minimise” your big product idea? Hit reply and give us the non-NDA version and we’ll brainstorm it a bit with you…

THE THREAD

Ok, so we’ve discussed what it means to do things that don’t scale. But how do you put that into practice? Ben Blaine shares some war stories from pioneering the concept in SA from the early days of Mxit, Snapscan, OfferZen and running the Investec Programmable Banking project.

Buckle up for storytime for your weekly dose of How Would You Build It on YouTube, Spotify or Apple Podcast.