I'll point out a few things about the algorithm selection process works at a high level. I work at another company dealing w/online ads.
Question: is your family member a CPC or CPM customer? I cannot claim familiarity w/how adwords works, but I see those as options when I try to make an acct (plus you see those 2 options available everywhere).
Also, I'm not sure what "hits per week" entails exactly. Is this some metric you can see on your personal adwords landing page? Any more descriptors of what it represents?
First, what happens when you visit a page that can show google ads? Roughly, Google will first evaluate what ad to show, if any. You can imagine this process as going over relevant ads (e.g., i'm searching for doctors; they won't even look at restaurants. also, they might not show ads from places that have been shown very often) and then scoring their relevance. This score can be based on things like what the user is doing now/recent history, quality of the business, where the user is located, time (of day, week, month, etc), and more. Then the best scoring ads are selected (or possibly nothing is shown if the scores are too low).
The scoring procedure isn't based on like, super intense math that I can sit down and do by hand or with a computer. Usually scores are composed of a bunch of different things and you add them together with weights (and these weights could vary depending on what kind of business it is). For example, maybe you think distance to provider is more important than time of day, so you increase the weight on the former. So choosing the weights and choosing what factors are important isn't a precise science: you have to test things live.
How is the testing done? Google pioneered the use of a process called "A/B testing." Say A represents the current system, and B represents the system with somethings changed (maybe different weights, new scoring factors, etc). You run both of them live simultaneously, splitting your traffic. So maybe you give A 60% of traffic and B 40% of traffic. Then after a long enough time, you can see how A and B performed for relevant metrics (like how often people click on ads), and decide to accept B, reject B, or run more tests.
I can't promise you this is how Google does their thing, but it'll be similar. So based on all of that, there's a couple of things that could be happening to your family member but my main guesses:
1) Google rolled out a new algorithm. Using something like A/B testing, they saw that it *overall* improved adwords' performance. Say the average with A was 5% and the average with B is 6%. Those are averages--some do worse, some better. With A, maybe your family member saw 7% and with B, he sees 3.5% (these percentages don't mean anything; im just making up a metric). So the change was good on the whole, but in general, no company will try to guarantee that a change is better for every single user/provider (the data is WAY too sparse, forgetting about how expensive this would be).
2) Maybe other advertisers signed up and now there's just more competition for essentially the same user-space.
3) Google contracts w/other sites to let those sites show Google ads (instead of their own ads); then Google + other site split the revenue. Maybe your family member saw a substantial part of his hits come from one of Google's partners, and *they* changed something.
More specific examples could include:
1) Internal metrics changed and now Google has decided that they were previously showing your family member's ad too often and they think this is a more reasonable level.
2) Google's relevance assessment changed so that now they think your family member's ad/business is appropriate in a smaller number of situations.
3) Similarly, maybe the way keywords (both pos & neg) are handled changed; maybe things got more restrictive so that the number of cases Google thinks your family member's ad is valid to show has gone down.
4) They changed something that they know worsens some metrics (like the one your family member cares about) but they feel it was worth it for other reasons (like making them more money).
5) Some kind of testing process is still going on and your family member is in a group running a shittier algorithm (doubtful but possible).
6) The change he's observing is coincidental.
7) and on and on and on
So again, I don't work for Google nor do I know exactly how their system works. But I'm just trying to point out that it's a very, very complicated system and it's not always easy or possible to figure out precisely what change is causing problems.
Just because he noticed performance dropping w/this radius vs areas change doesn't mean that's the cause: there could very well have been other changes rolled out around that time/at the same time. Alternatively, it's not even obvious that covering the same area under the new system as the old system would produce identical results.
In short, it's *very* unlikely that you'll find someone who can precisely tell you "Google changed ____ and now ____ is worse." Even if you found a Google employee who was willing to violate NDA, I highly doubt they could tell you either.
And finally, if he's not happy with Google, there are other places to advertise too.