Agreed. Radi, if you'd like to eliminate the issue from your Moz report, you can noindex, follow those category pages.

We use Sprout Social here at Moz, and have been pretty happy with it .:)

I'd recommend opening in a new tab. Part of that is personal preference, but I can't see any case in which it's a good idea to move someone off your site in that situation.

Hi Malika! How'd it go? Did everything work out?

Yeah, we probably should.

Yeah, I'd have to say that a 404 would be far preferable. A 301 would be ideal, but it would take some bandwidth to redirect to the next-most-relevant page whenever a deal expires.

I have to say, GTAMP, I agree with all the above advice. As Jared notes, this isn't something that's going to wreck your traffic on its own. However, David is spot-on that it's certainly worth fixing if the cost isn't too high.

Hi David!

You're getting great advice from Andy and Patrick, but I wanted to add that a site with a spam score of 4 and a DA of 45 is really unlikely to cause you a problem. Rand's initial post on spam score may clarify a bit, but we don't generally flag a link as potentially harmful until it has a score of 5.

I can't guarantee that site is _not _causing harm, but a 45 DA is reasonably solid. Really consider whether you want to risk losing any potential benefits from that link before you get rid of it.

...wow. Yeah, you're not kidding. Everything from the domain to "/Link-Exchange" looks so spammy as to be almost comical. If someone showed me that site, I'd tell them it's spam without hesitating.

I'm starting to lean toward killing those links, definitely. It's a tough call, though. Or at least a strange one.

I agree with Eric. No one source is going to give you a full picture of your link profile. Generally, OSE is best for measuring the overall strength of a full link profile, as many low-authority sites aren't indexed.

Also, keep in mind that there are a _lot _ of reasons that DA can go down, many of which have nothing at all to do with your specific link profile. That's why we recommend using it to benchmark against competitors rather than as an absolute score. Rand goes into more detail about that here:

DA/PA Fluctuations: How to Interpret, Apply, & Understand These ML-Based Scores

Might've gone a little endorse-happy there, but the three of you hit all the points I was going to make.