4 ms·

How I know the person who wrote this doesn’t understand user acquisition. They give an example that their conversion rate on their website is (cue some heavy m

How I know the person who wrote this doesn’t understand user acquisition. They give an example that their conversion rate on their website is (cue some heavy math /s) 0.16 % so if they want to acquire X customers they need to bring in Y traffic. But of course anyone who has done even the most basic user acquisition knows that the way you achieve a visitor to your site is more important than the visitor. Customers who show up organically have a vastly higher conversion rate than the randos who show up from your ad, in fact many of them probably aren’t even going to be real so your conversion rate is going to nose dive as soon as you start using paid acquisition. Happens every time.

refulgentis 3y ago
This isn't very kind, in fact it's an outright personal attack. Worse, it misses the point, it's about making a quick back-of-the-napkin calculation as business consultant while making OKRs. The problem with adding segmentation for more accurate calculations is there's always one more person who can sneer it's inaccurate because there wasn't enough segments. This type of person tends to move on from marketing / data science quickly, real world constraints are very frustrating to them
- romafirst3 3y ago
  This is happening in the real world, not some imaginary world and so we must talk about real effects. As soon as you turn on a torrent of paid traffic conversion rates go to shit. So your old conversion rate doesn’t work anymore and is just a random number that is an upper bound on your new conversion rate.
edmundsauto 3y ago
I wouldn't necessarily say organic is always better than paid, although that's a reasonable generalization. I think of it in terms of the yield curve - organic tends to be high quality but doesn't always scale. You drop your efficiency in the marketing media mix as you scale to reach a ROAS that meets the business need.
Vagantem 3y ago
Hi there, author of article here - I definitely see your point. I might have been a bit to general here, it's like another poster here says, it's very much napkin math. Yes, "better leads" are ofc always better than randos - but these numbers are from of a made up business, not the company mentioned in the beginning so if we imagine that 90% of the users already comes form paid, then we might want to invest more in paid. And if organic, maybe we invest in content and SEO instead. Overall, when setting objectives it is still very important to have some numbers to target - you can even change these targets later if it turns out there is a better way of getting those 1066 customers.
- romafirst3 3y ago
  My point was that the model used was oversimplified so much that it was worthless and would fall at the first hurdle. Your conversion rate is a function of the traffic. You can’t change the traffic without impacting the conversion rate. I would estimate that you might need 10-1000x the traffic depending on the quality of the traffic in your paid campaign, so these aren’t little numbers these are huge oversights.
  - Vagantem 3y ago
    Agreed, but again it obv depends on the situation - and as this is a made up scenario, this might look very different for real companies. Maybe they want to focus on another conversion metric to get to 1066 customers altogether. I just put website visitors as it fits nicely with the strategies and examples I'm planning to talk about in following articles :)

refulgentis 3y ago

This isn't very kind, in fact it's an outright personal attack. Worse, it misses the point, it's about making a quick back-of-the-napkin calculation as business consultant while making OKRs. The problem with adding segmentation for more accurate calculations is there's always one more person who can sneer it's inaccurate because there wasn't enough segments. This type of person tends to move on from marketing / data science quickly, real world constraints are very frustrating to them

romafirst3 3y ago
This is happening in the real world, not some imaginary world and so we must talk about real effects. As soon as you turn on a torrent of paid traffic conversion rates go to shit. So your old conversion rate doesn’t work anymore and is just a random number that is an upper bound on your new conversion rate.

edmundsauto 3y ago

I wouldn't necessarily say organic is always better than paid, although that's a reasonable generalization. I think of it in terms of the yield curve - organic tends to be high quality but doesn't always scale. You drop your efficiency in the marketing media mix as you scale to reach a ROAS that meets the business need.

Vagantem 3y ago

Hi there, author of article here - I definitely see your point. I might have been a bit to general here, it's like another poster here says, it's very much napkin math. Yes, "better leads" are ofc always better than randos - but these numbers are from of a made up business, not the company mentioned in the beginning so if we imagine that 90% of the users already comes form paid, then we might want to invest more in paid. And if organic, maybe we invest in content and SEO instead. Overall, when setting objectives it is still very important to have some numbers to target - you can even change these targets later if it turns out there is a better way of getting those 1066 customers.

romafirst3 3y ago
My point was that the model used was oversimplified so much that it was worthless and would fall at the first hurdle. Your conversion rate is a function of the traffic. You can’t change the traffic without impacting the conversion rate. I would estimate that you might need 10-1000x the traffic depending on the quality of the traffic in your paid campaign, so these aren’t little numbers these are huge oversights.
- Vagantem 3y ago
  Agreed, but again it obv depends on the situation - and as this is a made up scenario, this might look very different for real companies. Maybe they want to focus on another conversion metric to get to 1066 customers altogether. I just put website visitors as it fits nicely with the strategies and examples I'm planning to talk about in following articles :)