There are three realistic ways to get a product built when you do not have a product team. Each is right in some circumstances and wasteful in others, and the choice is usually made on price alone, which is how founders end up with the wrong one.
The three models
The agency or pod
A team of three to five: product, design, engineering, often a delivery manager. Cape Town has several strong ones.
Good at: breadth and pace. Multiple workstreams at once, design and build in parallel, and the bench to absorb someone getting sick.
Costs: typically R400,000 upward per project, or a monthly pod rate.
Fails when: the work is smaller than the team. A pod is a fixed shape, and paying for four people to deliver what two could is not a discount you can negotiate. You are also a client of a firm, not a partner to a person: the senior who sold the work is rarely the one doing it three weeks in.
The offshore development team
An overseas team, usually per-hour or per-feature. You supply the spec.
Good at: cost per unit of code, by a wide margin. If your spec is settled, this is real money saved.
Costs: R35,000 to R120,000 for a typical first build.
Fails when: the spec is not settled, which for a first product it almost never is. Nobody in the engagement is paid to challenge the brief. You will get exactly what you asked for, which is only useful if what you asked for was right. Timezone and context gaps make the feedback loop slow at precisely the moment you need it fast.
The fractional product lead
One senior product person, part-time or project-based, owning scope and decisions. Engineering either from your team or a network they bring.
Good at: judgement per rand. The decisions that determine whether the product works, made by someone who has made them before. Cheap to start, quick to stop.
Costs: R150,000 to R400,000 for a fixed-scope build; monthly rates for ongoing work typically run 35 to 50% of a full-time senior salary.
Fails when: you need volume. One person plus a small team cannot ship four workstreams at once. If you genuinely need breadth now, this is the wrong shape and an agency is the right one.
Side by side
| Agency / pod | Offshore | Fractional lead | |
|---|---|---|---|
| Best for | Breadth, funded scale-ups | Settled specs, mechanical builds | Deciding what to build, first products |
| Typical cost | R400k+ | R35k–120k | R150k–400k |
| Time to start | 2–6 weeks | Days | Days |
| Challenges your brief | Sometimes | No | That is the job |
| Accountable person | A team | A project manager | One named person |
| Scales to breadth | Yes | Yes | No |
Choosing
You know exactly what to build, and it is mechanical. Offshore. Do not pay for judgement you do not need. Write the spec carefully, because you will get precisely it.
You have several workstreams and funding to match. Agency or pod. Breadth is a real thing to buy and one person cannot fake it.
You are not certain what to build, and that is the actual risk. Fractional lead. The money you save is scope you never build, and that saving is larger than any hourly rate difference.
You are a corporate innovation team. Usually fractional first for the decision, then an agency for scale once the direction is proven.
The expensive mistake
The common error is using the cheapest model to answer the most expensive question.
Offshore teams are inexpensive per unit of code and cannot tell you whether the code should exist. Using one to explore an unvalidated idea means paying build prices for research, then paying again when the research comes back negative. The saving on the rate is dwarfed by the cost of the wrong build.
The inverse error is rarer but real: hiring a pod for a two-person job, and paying team overhead for eight weeks of work that did not need a team.
Honest disclosure
SILC is the third model, so treat this as a description of the market rather than a neutral verdict. What I will say without hedging: I have sent work to agencies when the breadth genuinely warranted a team, and told founders to go offshore when their spec was settled and my involvement would only have added cost.
If you are unsure which shape fits, the intro call is free and I will tell you if the answer is not me.