THE AWKWARD
BIT AFTER
“USERS”.

The three figures above describe cumulative reach, average monthly activity and people who paid during a period. The filing defines a paying user as a user who has made at least one monetary transaction in a given period. Its account and device measures may count one person more than once. Read the definitions ↗

So I wouldn’t divide the nine-month payer count by an average month’s active audience and call the result a conversion rate. The windows differ. A useful conversion measure needs a defined starting group, a payment event and a period in which that group could convert.

There’s nothing wrong with publishing several audience measures. MiniMax gives the definitions. The problem would arrive in a shortened presentation that quietly drops them. “Users” is doing a suspicious amount of overtime by that point.

The latest update reports first-half revenue of US$116.6m, up 283.1%. Gross profit was US$20.8m, with research and development expense of US$296.9m. Company results ↗

I’d keep the distinction between earning revenue and funding the wider business in the main presentation. Otherwise “the product makes money” can accidentally become “the company funds itself”. That is a much bigger claim.

The chart below follows a dollar of revenue through reported cost of sales. R&D sits separately underneath. These are company-wide averages. They do not establish an individual product’s margin, cash runway or the return on the next dollar spent.

A young model company can choose to invest well ahead of current income. Training work may support products and revenue beyond this reporting period. That is a fair argument for spending. It still leaves a founder with something to explain: which part of the spending improves delivery economics, which part develops future capability, and what evidence would show the plan is working.

For a smaller company raising capital, I’d bring that explanation forward. A demo can be excellent while the cost of serving the next customer remains uncertain. Both can be true. The presentation has room for both.

Hong Kong’s fundraising ecosystem already talks about business models, financial resilience and investor preparation. HKSTP’s scale-up guidance covers those subjects. The useful next step is making them concrete enough for someone to question.

If I were helping a founder prepare, I’d start with a short rehearsal. Explain the product using an actual customer task. Point to the person who pays. Show what happened after the first purchase. Then walk through the cost of delivering it. I’d write down every answer that depends on “we think”. Those are candidates for the next test.

A service business might discover that the attractive margin excludes the founder’s delivery time. A software business might have a paid pilot that still needs substantial manual support. A hardware team may have customer interest while manufacturing yield remains untested. Those distinctions belong in the business model before someone draws a very confident line into next year.

For an early-stage venture, some answers will be experiments and estimates. I’d label them that way. The raise can fund a defined uncertainty: whether buyers renew, whether delivery costs fall, whether a repeatable sales route exists. “Growth” is a rather large drawer to put all of that in.