So I think "it depends" for that also.
Here's my experience.
Slides are easy. An ES-1 from NIkon and a 60mm Macro on a full frame camera, an off-camera flash or even white light, and you can copy slides as fast as you can blow the dust off and they come through very nicely. Slides were the easiest thing I did.
Negatives a more problematic. You can get something like the Nikon ES-2 which is more prepared for loose negatives; I used the ES-1 and made a carrier that would fit inside of it (after taking out the springs that held the slides). That's the option for shooting them with a camera to yield a negative image, which you have to invert. There are tools to do it (Negative Lab Pro is a good one) in the context of lightroom, or you can drop into Photoshop and do it there (tons of tutorials online for that process).
Or for negatives you can use a flatbed scanner and carrier, and something like Silverfast or Vuescan software. Most people, as far as I know, will let such software produce a positive image.
Or you can buy a specialized negative scanner; I was unable to tell how much value they add vs taking photos with a DSLR.
By the way, B&W negatives were pretty universally easy to do; no color casts involved, much less fading.
Both of these steps above produce a positive image that is somewhere between terrific and horrible; it depends a lot on how faded the negative is, and the tools you use. It's very hard to tell from looking at the negatives. I had some which were too faded to use, I had others I could work on quite a bit and make tolerable. The worst were ones that were not uniformly faded, e.g. one side of the negative was a completely different color than the other. OK, that wasn't the worst, but I threw out the ones that had spills and sticky stuff that had ruined the negative entirely. I, and my "help", were not always careful over the years.
Then there's prints -- these are really easy to scan on a flat bed scanner -- get on that has at least 300dpi, maybe a bit more, and can scan 16 bit TIFF's, and you are pretty much set.
Here's where there is a big grey area between what a service would do and what you can do. Each of the above paths yields an image that is flawed -- it may have color issues as simple as white balance, or where preferential fading has just made a mess. It may have scratches, dust, stains, it may be over/under exposed (and was it scanned the best to accommodate that; do you want to go back and try again). Unless you were really a great photographer AND archivist for all those years, you will have a lot of problem results. EVERY shot will potentially have some issue you want to fix.
A service will have a set of tools they use, they will likely adjust white balance either automatically or by eye. They may have additional services for various levels of restoration. You won't know how much effort you want to put in, frankly, until you see the result. It's hard to tell a service in advance what to do, which are worth the effort, which are not. And if you have prints, and can just decide "the print looks better" -- how do you tell a service to make that decision.
And then... there's organization. Putting in metadata, doing face tagging, captions, dates... this can be a black hole in which you can get lost for months. But this is also an interesting point in favor of doing it yourself. I continually would run across a set of images I did not recognize, and went into detective mode -- when it was it, who was it. If you are doing that as you go, context can matter -- notes on the sleeve, a date on the photo bag from the lab, which lab it was. Or maybe you were organized and all this is easy. If so please do not tell me!
I spent years doing this a few weeks at a time in spurts, but 90% of that was in restoration or organization, not actually digitizing. Once I got the ability to take photos of negatives and slides, and a good flatbed scanner (I think $200? Not too expensive) that was pretty quick. but you can spend an infinite amount of time retouching, and a lot of time finding out "who is that blond with Cousin Charlie in 1943".
I found it very convenient to be making these decisions -- including to throw out hundreds of images -- as I went. I'd start scanning in a bunch of images and realize they were snapshots of (say) zoo animals from visiting washington DC in high school. Not people, not relatives, not art -- I just though a zebra was interesting because it was probably the first time I saw one. 99% of those I just threw away. But I didn't decide that until I saw a few positives and remembered what they were. Having your brain in the loop as they get scanned can add a lot of value.
Anyway... I ramble. It's a hugely subjective decision, how much effort you are willing to put in. An aspect in favor of a service is you can just send them ALL off and other than paying the bill, they ALL come back done and you do not have it hanging over you for ages "should I go back and work on it some more". But if you are a decent photographer/post-processor already, I almost guarantee you will not be as happy with the results as if you did it yourself. IF you actually finish doing it yourself. Took me a LONG time.
