I back things. I like backing things. Somebody has an idea that no company would greenlight, a few thousand strangers decide it should exist, and sometimes it does. When it works it is the best thing about this industry, and I want it to work every time.
It does not work every time, which is why I read these pages the way I do. Here is the order I go in.
The team comes first, before the product. Have these people shipped anything? A hardware background, a previous campaign that delivered, a founder with a name attached to earlier work: any of it changes the odds enormously. First-time teams do succeed, and I have backed plenty, but a first project and a complicated product together is a lot of weight for one team to carry.
Next, the prototype, and specifically whether I am watching one. A render is a drawing. An animation is a drawing that moves. What I want is a continuous shot of a real object being used by a person, with no cut at the moment the interesting thing happens. The cut is the tell. If the only footage of the mechanism working is three seconds long and edited on both sides, that mechanism is probably still a problem.
Then the risks and challenges section, which is required on every Kickstarter page and which most people skip. A good one is specific and slightly uncomfortable: it names the injection molding, the certification, the firmware, the supplier who might be late. A bad one says the team is confident and experienced. I have come to trust a campaign more when its risks section worries me, because it means somebody has actually looked.
Then the dates. A delivery date on a campaign page is a hope in a nice font. What I check instead is the gap between the end of funding and the first promised shipment. Under six months for a new physical product is optimistic unless tooling already exists, and if the page says tooling is done, they will usually show you the tool.
Stretch goals need reading sideways. A stretch goal that adds a color is fine. One that adds a feature, a second model or an app is a new project bolted onto a schedule that was already tight, and it is one of the most reliable causes of a year of delay. I like a campaign that refuses to add them.
Last, understand the deal you are making. Kickstarter is not a shop. Backing is not ordering, your card is charged when the campaign ends rather than when anything ships, and the platform's terms put the obligation to deliver, or to refund what is undelivered, on the creator rather than on Kickstarter. The platform runs the pledge. The team owes you the object.
So: have they built anything before, is there an unedited prototype, does the risks section sound like somebody who has been in a factory, is the schedule sane, and can you afford for it to go wrong?
The short version: back it with money you are willing to turn into a story about a thing that did not arrive, and then go in hoping, because most of the good ones started exactly here.
