Why a photograph is rarely an acceptable submission
Almost nobody wants a JPEG of a document. Application portals, HR departments, banks and universities ask for PDFs because a PDF has fixed page dimensions, prints predictably, holds several pages in one file, and cannot be casually edited in the way an image can. A photograph of your degree certificate is the right information in the wrong container.
The practical consequence is a step almost everybody has to take at some point: photograph or scan something, then turn it into a PDF before anyone will accept it. Doing that on your own device rather than uploading to a converter matters more here than almost anywhere else on this site, because the images people convert are overwhelmingly identity documents — passports, licences, certificates, signed forms, bank paperwork.
It is also the fix for a subtler problem. Sending five separate photographs means the recipient has to open five files and work out their order. One PDF with the pages in sequence is a single attachment that reads the way you intended.
Page size, orientation and margins
Each image is placed on its own page, scaled to fit while keeping its aspect ratio, so nothing is stretched or cropped. A photograph taken in portrait produces a portrait page and a landscape one produces a landscape page, which is normally what you want and occasionally not — a document photographed sideways will produce a sideways page, and the fix is to rotate the resulting PDF rather than to fight the converter.
Aspect ratio is preserved rather than forced to the page, which means a very wide or very tall image leaves white space on the page's short axis. That is the correct trade: distorting a document to fill a page makes it look wrong and, on anything with a signature or a photograph, can make it look tampered with.
If the recipient specified a page size — A4 is the usual requirement outside North America, Letter inside it — check the result rather than assuming. A mixed set of photographs can produce a document whose pages are all slightly different, and some strict upload systems object.
Getting a usable result from a phone photograph
Most rejected document photographs fail for the same handful of reasons, and all of them are easier to fix before conversion than after. Shoot in the brightest even light you can find and avoid direct overhead light, which throws a shadow of your own phone across the page. Put the document on a plain contrasting surface so its edges are unambiguous. Hold the camera parallel to the page rather than at an angle, because keystoned text is the single most common reason a document is sent back.
Fill the frame with the document, but leave a small margin so no edge is clipped — a certificate with a corner missing looks like a document with something to hide. Check focus on the small print, not the large heading, since it is the small print that becomes unreadable first.
If the finished PDF is too large for the form you are filling in, convert first and compress afterwards rather than reducing the photographs beforehand. Compressing once at the end gives a better result than degrading each image separately and then assembling them.