4 ms·

I am genuinely curious if people here on HN calculate monetary value attached to their website visitors. I know patio101 (from BCC) must be doing that but does

I am genuinely curious if people here on HN calculate monetary value attached to their website visitors. I know patio101 (from BCC) must be doing that but does anybody else here does that?

zaidf 17y ago
It's very challenging for ad-supported models especially during start-up stage where there is a lot of fluctuation in ad rates. It can become easier to calculate after a couples years when your company's settled down and you've some consistent ad rev sources. I think a lot more folks that sell a service get into this kind of calculation from the get go. Isn't this what conversion rates/ROI are all about?
- paraschopra 17y ago
  Yep, it is mostly about conversion rates. But this metric includes retention rate, which almost every website (including ad supported ones) should optimize for. I agree with you that for websites just starting up, it can be hard. In that case before full knowledge is there, industry standard conversion rates and value of goals can be plugged in. Replace those estimates by actual values as soon as you have a good idea about rev sources, as you say.
patio11 17y ago
I have done this before, and even encouraged it before. Mea maxima culpa. I am sorry and will not repeat these numbers again. Here's my thinking: At a point in the past I made $80 for every 1k visitors to my website. In October, I made $72 for every 1k visitors to my website. (Some folks would call these figures eCPMs.) Now we come to the crucial question to ask about any metric: does this provide actionable insights? Can you say, looking at those numbers, that I was doing better in the past than I did in October? Well, I suppose you could, but you'd be wrong: I've quadrupled the bottom line since then. The difference in the two eCPMs is driven largely by the effectiveness of my Halloween promotions, which brought in lots of paying customers mixed in with a larger-than-usual horde of folks who did not happen to pay me money. Incidentally, I totally agree with you with regards to page views and visitor numbers: meaningless. Incidentally: aggregating your statistics over multiple sources of traffic tends to blunt their utility. I tend to watch conversion rates like a freaking hawk, for example. However, even these are fairly useless as aggregate numbers. For example, consider that I have two sources of traffic which convert, a priori, at 10% and 20%. My traffic mix is 50:50. This will produce a sitewide conversion rate of 15%. Now increase the absolute count of visitors from the first source by 20% (maybe I did well at SEO this month), and my sitewide conversion rate falls to 14.54%. Looking at the conversion metric tells you you're doing something wrong, looking at your wallet tells you you're doing something right. The wallet wins. This isn't just a theoretical issue, it comes up at the Business of Software boards all the time, and causes needless anxiety among software developers who know just enough about optimization to hurt themselves. Edited to continue: So what do I think about tracking LTV? Well, subject it to the metrics acid test: does it give you actionable insights? I think sitewide LTV does not give you actionable insights. Sometimes LTV going up is a good thing. Sometimes LTV going down is a good thing. (Many startups here will see LTV plummet at a point in their lifecycle. It is colloquially known as "crossing the chasm".) However, with a sufficiently advanced testing system, you could probably use LTV effectively. For example, if you could split test customers across a cohort and compare LTV for the groups versus the conditions you exposed them to, that might be worthwhile information. (Example: suppose you're Joe Random Subscription Billing Service. You could create a cohort out of everyone signed up in the last three days, then randomly apportion them into two groups. Group A will receive no treatment. Group B will receive an email right before their next cycle offering them a 50% discount if they upgrade to the next service level. Six months from now you measure their LTVs. Those LTVs could, potentially, be meaningful.)
- paraschopra 17y ago
  Well if you ask me, you did little worse in October than the past. And if the trend in whatever money you make per 1k of visitors is consistently downward, I will be really worried. While it may not provide actionable insights (that is what bounce rate and conversion rates are for), I don't see why you should not be concerned about LTV at all. What you are saying is that one cannot increase visitor value while increasing total traffic? Hence arent concerned about per visitor value as long as your bottomline is quadrupled. Do you think your bottomline would have swollen even more if $80/k visits didn't drop to $72/k visits. Or you think it wasn't possible because traffic increased and one cannot maintain the visitor value as before? Having an eye at individual conversion rates is better than having an aggregate conversion rates. Definitely I agree with you on this. However, people tend to cling on to aggregate numbers because brain can hold limited information at a time. Better to have a vague idea of what is happening than no idea at all :)
  - patio11 17y ago
    I disagree, in the strongest possible terms, that bounce rate is a useful metric. Somewhat surprisingly to me, it seems that folks I find very credible in the field disagree with me. http://www.kaushik.net/avinash/2007/08/standard-metrics-revisited-3-bounce-rate.html http://www.kaushik.net/avinash/2007/08/standard-metrics-revi... He knows what he's talking about, and he says "Bounce rate is a beautiful way to measure the quality of traffic coming to your website", which do not appear to include the words "of less value than the toenail clippings of squirrels", which is about where I think bounce rate should be. In fact, I think it insults squirrels that I would mention their toenail clippings in the same sentence as bounce rate. The gap between what I believe and what everyone else believes is so large that this suggests I should make the discussion a blog post. I'll do that later.
    - paraschopra 17y ago
      Hmm.. your point looks somewhat surprising (and confusing) to me. (Where did squirrels enter the picture). Will be definitely looking forward to your post.

I am genuinely curious if people here on HN calculate monetary value attached to their website visitors. I know patio101 (from BCC) must be doing that but does anybody else here does that?

zaidf 17y ago

It's very challenging for ad-supported models especially during start-up stage where there is a lot of fluctuation in ad rates. It can become easier to calculate after a couples years when your company's settled down and you've some consistent ad rev sources. I think a lot more folks that sell a service get into this kind of calculation from the get go. Isn't this what conversion rates/ROI are all about?

paraschopra 17y ago
Yep, it is mostly about conversion rates. But this metric includes retention rate, which almost every website (including ad supported ones) should optimize for. I agree with you that for websites just starting up, it can be hard. In that case before full knowledge is there, industry standard conversion rates and value of goals can be plugged in. Replace those estimates by actual values as soon as you have a good idea about rev sources, as you say.

patio11 17y ago

I have done this before, and even encouraged it before. Mea maxima culpa. I am sorry and will not repeat these numbers again. Here's my thinking: At a point in the past I made $80 for every 1k visitors to my website. In October, I made $72 for every 1k visitors to my website. (Some folks would call these figures eCPMs.) Now we come to the crucial question to ask about any metric: does this provide actionable insights? Can you say, looking at those numbers, that I was doing better in the past than I did in October? Well, I suppose you could, but you'd be wrong: I've quadrupled the bottom line since then. The difference in the two eCPMs is driven largely by the effectiveness of my Halloween promotions, which brought in lots of paying customers mixed in with a larger-than-usual horde of folks who did not happen to pay me money. Incidentally, I totally agree with you with regards to page views and visitor numbers: meaningless. Incidentally: aggregating your statistics over multiple sources of traffic tends to blunt their utility. I tend to watch conversion rates like a freaking hawk, for example. However, even these are fairly useless as aggregate numbers. For example, consider that I have two sources of traffic which convert, a priori, at 10% and 20%. My traffic mix is 50:50. This will produce a sitewide conversion rate of 15%. Now increase the absolute count of visitors from the first source by 20% (maybe I did well at SEO this month), and my sitewide conversion rate falls to 14.54%. Looking at the conversion metric tells you you're doing something wrong, looking at your wallet tells you you're doing something right. The wallet wins. This isn't just a theoretical issue, it comes up at the Business of Software boards all the time, and causes needless anxiety among software developers who know just enough about optimization to hurt themselves. Edited to continue: So what do I think about tracking LTV? Well, subject it to the metrics acid test: does it give you actionable insights? I think sitewide LTV does not give you actionable insights. Sometimes LTV going up is a good thing. Sometimes LTV going down is a good thing. (Many startups here will see LTV plummet at a point in their lifecycle. It is colloquially known as "crossing the chasm".) However, with a sufficiently advanced testing system, you could probably use LTV effectively. For example, if you could split test customers across a cohort and compare LTV for the groups versus the conditions you exposed them to, that might be worthwhile information. (Example: suppose you're Joe Random Subscription Billing Service. You could create a cohort out of everyone signed up in the last three days, then randomly apportion them into two groups. Group A will receive no treatment. Group B will receive an email right before their next cycle offering them a 50% discount if they upgrade to the next service level. Six months from now you measure their LTVs. Those LTVs could, potentially, be meaningful.)

paraschopra 17y ago
Well if you ask me, you did little worse in October than the past. And if the trend in whatever money you make per 1k of visitors is consistently downward, I will be really worried. While it may not provide actionable insights (that is what bounce rate and conversion rates are for), I don't see why you should not be concerned about LTV at all. What you are saying is that one cannot increase visitor value while increasing total traffic? Hence arent concerned about per visitor value as long as your bottomline is quadrupled. Do you think your bottomline would have swollen even more if $80/k visits didn't drop to $72/k visits. Or you think it wasn't possible because traffic increased and one cannot maintain the visitor value as before? Having an eye at individual conversion rates is better than having an aggregate conversion rates. Definitely I agree with you on this. However, people tend to cling on to aggregate numbers because brain can hold limited information at a time. Better to have a vague idea of what is happening than no idea at all :)
- patio11 17y ago
  I disagree, in the strongest possible terms, that bounce rate is a useful metric. Somewhat surprisingly to me, it seems that folks I find very credible in the field disagree with me. http://www.kaushik.net/avinash/2007/08/standard-metrics-revisited-3-bounce-rate.html http://www.kaushik.net/avinash/2007/08/standard-metrics-revi... He knows what he's talking about, and he says "Bounce rate is a beautiful way to measure the quality of traffic coming to your website", which do not appear to include the words "of less value than the toenail clippings of squirrels", which is about where I think bounce rate should be. In fact, I think it insults squirrels that I would mention their toenail clippings in the same sentence as bounce rate. The gap between what I believe and what everyone else believes is so large that this suggests I should make the discussion a blog post. I'll do that later.
  - paraschopra 17y ago
    Hmm.. your point looks somewhat surprising (and confusing) to me. (Where did squirrels enter the picture). Will be definitely looking forward to your post.