3 ms·

I agree it’s dumb but maybe you overestimate the masses. Maybe on some level it works for the execs, especially in the short term.

jerf 4y ago
That's what powvans is referring to with "brilliant way to drive conversion". Measure "engagement" by "how many songs the user starts playing" and combine that with "clicking on a song the first time actually plays a different song" and you'll get more clicks to start songs as the user tries to correct their problems. This isn't a good thing, the metric has been over-optimized, but if you're paid by metrics this is what happens. Likely there is a programmer or two well aware of the issue and the presumably 1-10 line fix, but has either tried it and been told to revert it, or simply knows they can't push it, because they know what it would do to the metrics. Probably a good number of the complaints people are leveling here against Amazon services source from that. For instance, mixing paid options into the list as quickly as possible when you explicitly asked for free stuff can't hardly help but "convert better" than letting the user truly browse through the free stuff; if so much as one person ever gives in and buys something that way, or even browses, leaves, comes back to the same screen later and forgot they were supposedly looking at free stuff and buys something, boom, better conversion rate than pure free. Doesn't take much to move the needle off of "zero". But it's still not what we asked for. (& yes, I know that it's non-free stuff in the recommendations when viewing a free option, not unfree stuff mixed right into free carousel, but I think it's fair to say customers are definitely interested in a "free only right now please" view and a good percentage of them probably think that's what they're getting when they're in the "free" tabs.)

3 ms·

I agree it’s dumb but maybe you overestimate the masses. Maybe on some level it works for the execs, especially in the short term.

by tartoran 4y ago

jerf 4y ago
That's what powvans is referring to with "brilliant way to drive conversion". Measure "engagement" by "how many songs the user starts playing" and combine that with "clicking on a song the first time actually plays a different song" and you'll get more clicks to start songs as the user tries to correct their problems. This isn't a good thing, the metric has been over-optimized, but if you're paid by metrics this is what happens. Likely there is a programmer or two well aware of the issue and the presumably 1-10 line fix, but has either tried it and been told to revert it, or simply knows they can't push it, because they know what it would do to the metrics. Probably a good number of the complaints people are leveling here against Amazon services source from that. For instance, mixing paid options into the list as quickly as possible when you explicitly asked for free stuff can't hardly help but "convert better" than letting the user truly browse through the free stuff; if so much as one person ever gives in and buys something that way, or even browses, leaves, comes back to the same screen later and forgot they were supposedly looking at free stuff and buys something, boom, better conversion rate than pure free. Doesn't take much to move the needle off of "zero". But it's still not what we asked for. (& yes, I know that it's non-free stuff in the recommendations when viewing a free option, not unfree stuff mixed right into free carousel, but I think it's fair to say customers are definitely interested in a "free only right now please" view and a good percentage of them probably think that's what they're getting when they're in the "free" tabs.)