Platform proposal · NW-AAA-2026-001
Nominees.
Voting.
A result that
holds up.
The awards platform for the Amapiano Africa Awards. Built on an engine already running a national South African vote.
Platform proposal · NW-AAA-2026-001
The awards platform for the Amapiano Africa Awards. Built on an engine already running a national South African vote.
The deadline that binds is not 14 November. It is 13 October, the day the public vote has to be live.
A public vote needs roughly four weeks in market to produce a number worth announcing, and the result has to be closed and checked before anyone walks on stage.
Work backwards from 14 November and voting opens in the middle of October. That is about six weeks from now. Everything on this page is built from that date, not from the ceremony.
Every week this waits comes off the voting window, not off the build. A short voting window produces a smaller number, and the number is what you sell to a sponsor.
Thirty categories across the Vocalist, DJ, Dance and Producer segments. The nominee roster. The public vote. And the admin that lets you run all of it without phoning us.
Any designer can build nominee pages and a voting form. What that does not survive is the morning after, when a losing artist's camp says in public that the vote was bought. In this category that is not a risk. It is a pattern, and it has followed South African music awards for years.
So the money goes on the counting, not the pages.
The limit is enforced inside the database, not in the browser. Two taps landing in the same instant cannot both slip past it. A limit enforced in the browser is not a limit, it is a suggestion.
Verification tied to something that costs a person effort to duplicate, rather than a free email address that can be generated a thousand at a time.
Every screen shows both. Five hundred votes from ten people and five hundred from five hundred people are not the same result, and only one of them can be published.
Nothing is overwritten, and tallies are snapshotted every hour. If the outcome is questioned in December, that is how it gets answered.
The rules page states the limit, the window and what invalidates a vote, before voting opens. Publishing the method in advance is the cheapest defence there is.
Burst clusters reviewed, rate limits tuned against real traffic rather than guesses, and the ability to invalidate a vote with the reason recorded against it.
None of this is discoverable in six weeks by someone building it for the first time. It is the part you are actually buying.
Lumanaire built and runs the Content Creator Awards, live now at contentcreatorawards.co.za. It is in its fourth year, Samsung sponsor it, and its public vote is open as you read this.
Open it. Look at the nominee pages, the category pages, the voting flow. Cast a vote. This is not a portfolio screenshot, it is a running system you can inspect before you commit anything to us.
That is a national online audience, not a headcount at a door. Five thousand people in a dome will never generate that load, because voting traffic has nothing to do with venue capacity. The figures above are the online channel only and exclude votes cast by SMS on that campaign, which a third party counted and which never reached our system, so the true combined total was higher by an amount we cannot state.
Pretending otherwise would waste your time, so this sits here at the front rather than buried in the small print.
Basadi runs SMS to a short code at R1.50 a vote and sells bundles from R15. SAMA runs USSD at the same rate. Those shows earn revenue from voting. A meaningful part of your audience is on prepaid, and on a Tuesday they may not have data.
The recommendation is both channels, for different reasons. Free web voting gives you scale, a clean first-party database and a result nobody can characterise as sold. A paid SMS or USSD channel reaches the audience web voting misses, and it is a revenue line you own outright. The published method then states how the two combine.
That is deliberate rather than modest. A revenue share would align us with vote volume, and what you are buying is vote integrity. Those two pull in opposite directions and we would rather be paid a fee and sit on the right side of it. The SMS revenue is yours.
The aggregator bills the sender and sets the per-message rate, so that relationship sits with you and nothing is marked up by us. Our side is the integration and the reconciliation. Worth starting that conversation this week, because provisioning a short code has its own lead time and it is the one item here we cannot compress on your behalf.
Your partnership deck asks a title partner for R5 million over five years, or R6.5 million over three. Read what it promises them: logo placement on all produced assets including the website and newsletters, inclusion in emails and SMS communication to databases, branded content, ticket giveaways and product sampling.
Most of that list needs a platform to exist before it can be delivered. There is no newsletter without a database, no email to a database nobody has built, and no logo on a website that has not been made.
It is also the evidence. A brand deciding whether to back a first ceremony is buying access to an audience, and audiences are proved with numbers. Sixty days of live public voting produces a list of real people with a name, a contact record and a recorded consent. That is a very different conversation to have with a sponsor than a deck is.
A voting platform cannot manufacture an audience out of nothing. Votes come from promotion: radio, the artists themselves, media partners and social. We build the machine that captures and counts, and we will make it easy to share, but the traffic is driven by your marketing and your talent. If the campaign is quiet the number will be quiet, and that will not be a platform failure.
Under POPIA, a marketing message to someone who did not opt in is not a marketing problem, it is a legal one, and consent cannot be added retrospectively to records already collected. So it is captured at the point of voting from the very first vote, with the wording and version stored against each record. It costs almost nothing at the start and cannot be fixed afterwards. Without it, the database you paid to build is not usable for the exact promise your sponsorship deck makes. Sponsors receive sends, not the list.
By Tuesday 9 September
Approval, and one answer
Whether the thirty awards are decided by public vote, by a jury, or by a combination of both. It changes the build materially and it is the one thing needed before the plan is fixed.
Week of 8 September
Kickoff
Brand assets and domain control needed from Nova World.
By 16 September
Final categories and segments
On or about 26 September
Public site live
A real URL on your own domain that you can put in front of a sponsor, with the partner wall already on it.
By 30 September
Nominee roster and cleared images
The row that moves every row after it. Late delivery shortens the voting window rather than moving the ceremony.
Week of 6 October
Voting engine complete
Rules published, load test passed against a target agreed in advance.
On or about 13 October
Nominees announced, voting opens
On or about 7 November
Voting closes, tallies frozen
Integrity report issued to Nova World.
13 to 15 November
Show weekend
Named standby with a direct escalation number across all three days, included rather than sold as an extra. Winners published on the night.
Written here rather than waiting to be asked, because event dates move and it is better settled in September than argued about in November. Every stage is earned on delivery of software, not on the ceremony taking place, so a postponed show does not un-deliver a website. Stages already delivered stay invoiced, stages not started are not owed, and remaining dates are re-agreed in writing against the new date with no penalty for the move. Dates inside the platform are configuration rather than a rebuild, so shifting them within the same cycle sits inside the monthly at no extra cost. And if Nova World stops at any point, the voter database collected up to that moment is exported to you in full, at no charge. Your data is never held against a fee.
Lumanaire is not a VAT vendor, so nothing is added. R75,000 is the whole amount payable on the build, and it is spread across three equal stages rather than a single deposit, so it lands alongside your other commitments instead of ahead of them. Each stage is invoiced before that block of work begins, and each one opens the next.
| Stage | What it opens | Amount |
|---|---|---|
| On acceptance | Kickoff, branding, categories, and the public site through to launch | R 25 000 |
| Before Phase 2, on or about 29 September | Nominee roster, voting engine, integrity tooling and load testing | R 25 000 |
| Before voting opens, on or about 13 October | Public voting goes live and the campaign period begins | R 25 000 |
| Total | No VAT is charged | R 75 000 |
Your first payment produces a live public site on your own domain before the second payment falls due. That is deliberate. It means you have something to put in front of a sponsor while the rest is still being built.
During the campaign this is an operations job rather than a hosting line. While voting is open the vote is watched daily, rate limits are tuned against real traffic, burst clusters are reviewed, nominee corrections and withdrawals are handled the same day, campaign email goes out to the voter database, and sponsor placements are checked and reported. Underneath it sits secure hosting, ongoing maintenance, performance work, backups, security updates and monitoring. It begins at go-live and is invoiced monthly in advance. Nothing monthly is payable during the build, and it can be cancelled on thirty days' notice with your full data set exported to you.
Priced separately so the core platform stays where it is. Each can be added at any point before the vote opens.
Integration and reconciliation so votes cast on a short code land in the same tally as web votes and show on the same dashboard. The aggregator bills you directly at their per-message rate with nothing added by us. Quoted firm the day you name the provider.
A language toggle across navigation, voting instructions, rules and FAQs in up to three languages alongside English. Nominee biographies and news stay in English unless translated copy is supplied. The rules page is the one place multiple languages genuinely matter, because fairness has to be readable.
Scanning and guest management across the three weekend events with media, artist and crew tiers. This depends entirely on who sells the tickets, because it has to reconcile against their system.
Your company site is separate work, quoted on its own so the two do not hold each other up. This proposal is the awards platform only.
Thirty categories at five nominees each is one hundred and fifty artist records. Each needs a legal name, a stage name, a category, a handle, a short biography and a press image you have the right to publish. The platform is deliberately built so that this arriving late does not stop the site going live: categories, dates, the programme, the sponsor wall and interest capture are all configuration and can be public without a single nominee name. A slip on the roster does not by itself trigger a payment. Stage payments are earned on our delivery, not on your calendar.
Nothing here covers event management, ticketing, staffing, on-site production or attendance beyond the standby described above.
None of these block approval or the start of the build. All of them need answers before the vote opens, and the first needs an answer before the plan is fixed.
The ceremony is on 14 November, but the deadline that actually binds is the middle of October, when voting has to be open. That is six weeks. Every week this waits takes a week off the voting window rather than off the build.
Reply on either channel and we send the first stage invoice and start the same week.
Proposal reference NW-AAA-2026-001. Valid until 9 September 2026.
