No company truly wants its customers to be made aware of recalls for its products. A business has very little to gain (some goodwill), but much to lose (a lot of money). They put in the minimum effort mandated by law, and have no incentive to open themselves up to increased recall numbers. Companies are not going to willingly provide APIs that allow third party companies to simplify or centralize registrations.
Thus, a company wanting to do this would be left performing manual data entry or automated scraping of existing sources. It would also be nice to notify users of products which have multiple reports of a defect, before recalls are put out. However I suspect companies would be quick to sue over unsubstantiated claims.
Monetization of such a product would be limited. Very few people would pay for this service. You're essentially left with building a profile based on a user's purchased products, used to sell ad space and/or push affiliate links. With how infrequently a user would be opening the website or mobile app, I can't imagine the income being enough to sustain even a small startup. Assuming there are existing public catalogues of recall information that can be scraped, I could see a solo developer doing this as a quick project for some side income. I don't see this as being a full startup due to, as you put it, "vendor compliance".
I just thought about some of the recall info sheets I've seen at the local supermarket: the cable entry on an electric blanket was misdesigned and had the potential to catch fire (:D), certain food items unknowingly contained allergens (presumably due to insufficient communication with manufacturers), the packaging (removable twist top) of a package of baby food could disintegrate (small pieces of the safety seal could break off) and present a choking hazard, etc.
These examples of recalls all share the element of preventing further injury and minimizing the chances of some disaster occurring. Hmm, I see what you mean by the goodwill aspect now.
As for sourcing data and payment, I was actually envisaging a two-sided product - creating the app described before for consumers, but also providing retailers and manufacturers with business-specific portals that allowed them to manage all aspects of a recall internally. I was thinking along the lines of "you store your data in the cloud so you don't have to maintain the associated infrastructure; this is the same, let us understand how you manage your recall processes so we can add this knowledge to a mega-system we build for you that will be ten times as good as what you've already got, and then just pay us $lessthanexpected to maintain it". Essentially solicited outsourcing.
With that done, that solves the industry integration problem (oh, and keeping the lights on); once vendors hit "publish" I'd have my customer-specific data.
Besides B2C this could work for B2B recalls of industrial and commercial equipment too. There could be pricing tiers to manage configurations of fleet or other large volumes of business-related equipment too.
Also, while pondering the idea earlier, I realized I could also expand horizontally to include warranty - it seems relevant, and I'm collecting product info anyway. Although about two seconds after the "that would be really cool" little red alarms started going off as I thought about all the spam (in the form of fraudulent claims) that I'd need to figure out a good way to filter.
I was wondering about ways to benefit (conscientiously, if you'd believe it) from customers' product profiles as well. FB (as one example) loves data, but I don't know what their appetite is like in terms of data volume or depth, and whether I'd be comfortable with what that industry would want.
Lastly, I envisaged the app as "install and forget" - maybe even going so far on Android as to hide the app icon completely (if that's deemed polite?). No bandwidth is used (with background messaging etc) until a recall occurs - only then does it spring into action. (Maybe if warranty stuff is factored in it could alert when devices are about to go out of warranty too.)
Thus, a company wanting to do this would be left performing manual data entry or automated scraping of existing sources. It would also be nice to notify users of products which have multiple reports of a defect, before recalls are put out. However I suspect companies would be quick to sue over unsubstantiated claims.
Monetization of such a product would be limited. Very few people would pay for this service. You're essentially left with building a profile based on a user's purchased products, used to sell ad space and/or push affiliate links. With how infrequently a user would be opening the website or mobile app, I can't imagine the income being enough to sustain even a small startup. Assuming there are existing public catalogues of recall information that can be scraped, I could see a solo developer doing this as a quick project for some side income. I don't see this as being a full startup due to, as you put it, "vendor compliance".