4 ms·

Conversion Optimizer: AdWords, Done Right: MicroISV on a Shoestring

25 points by rfreytag 17y ago

patio11 17y ago
Wow, a blog post of mine from late 2007. Wasn't expecting that on HN this morning. Here's the update for 2009: everything I said earlier remains true to a first approximation. AdWords is far and away the largest expense for my business and it also has the most predictable ROI of anything I do. Day in, day out, without my intervention it continues making me money. I really can't bang on that enough -- in 2007 prior to Conversion Optimizer I was having to log into AdWords and adjust daily, now the thing just runs on autopilot and I check back, oh, a little less than once a month? If you've got questions, fire away.
- gfodor 17y ago
  How hard was it to rig up your conversion tracking to Google? Was there any fear you had putting such a thing on auto-pilot with your wallet? Any ways you would have wanted it to be easier? Are there changes you'd make that might improve your ROI?
  - patio11 17y ago
    Conversion tracking: It is close to trivial to get a basic level of conversion tracking established -- copy/paste some Javascript into a page which probably already exists in your conversion funnel. Back in 2007 Google's conversion tracking had some accuracy issues (i.e. it was nowhere close to what my numbers were -- counting 50% of transactions whereas I expected 90% of customers to see the page I was counting at). That has since cleared up. Fear of autopilot: They have pretty comprehensive budgeting control (you set it in terms of a per-day thing, but that is basically a marketing-friendly wrapper around a per-month budget). That lets you sharply limit the risk -- I think at the start I had a daily budget of $10, which means if the process went out of control for a month I'd be out at most $300 when I got the next bill and terminated it. They've since performed well enough that I've adjusted my budget upwards several times until they were unable to spend most of it. (The default for AdWords is to adjust your bidding to spend up to your budget -- for example, if you have $500 unspent, they'll adjust your bids upwards to get you more clicks and spend that last $500. This default is in Google's interests, but not in yours, because it will result in you paying absolutely stupid per-conversion costs at the end.) For this Halloween I'd feel comfortable authorizing them to spend several thousand dollars. (Halloween is my best season and its the only one where my ads ever get budget-constrained.) Ways it could have been easier: Conversion Optimizer is about as easy as I could want it to be. AdWords, hmm... I'd like there to be a clone of me who I can get to write creatives (that is advertiser-speak for ads and landing pages) and check if they're performing well without me having to actually do it. Everything about quality score could be more transparent, particularly for newbies to the system. Changes to improve ROI: I've always got ideas for this, I just don't have nearly enough time to implement and rigorously test them all. My big project right now is turning my desktop application into a web app, which I think will result in a substantial across-the-board increase in conversions due to the elimination of the download, install, execute cycle and the possibility to collect email addresses to implement an autoresponder & etc.
    - gfodor 17y ago
      thanks for the reply -- do me a favor, if you have time, I'd like to chat more offline, email me at gfodor at gmail if you're up for it!

25 points by rfreytag 17y ago

Wow, a blog post of mine from late 2007. Wasn't expecting that on HN this morning. Here's the update for 2009: everything I said earlier remains true to a first approximation. AdWords is far and away the largest expense for my business and it also has the most predictable ROI of anything I do. Day in, day out, without my intervention it continues making me money. I really can't bang on that enough -- in 2007 prior to Conversion Optimizer I was having to log into AdWords and adjust daily, now the thing just runs on autopilot and I check back, oh, a little less than once a month? If you've got questions, fire away.

gfodor 17y ago
How hard was it to rig up your conversion tracking to Google? Was there any fear you had putting such a thing on auto-pilot with your wallet? Any ways you would have wanted it to be easier? Are there changes you'd make that might improve your ROI?
- patio11 17y ago
  Conversion tracking: It is close to trivial to get a basic level of conversion tracking established -- copy/paste some Javascript into a page which probably already exists in your conversion funnel. Back in 2007 Google's conversion tracking had some accuracy issues (i.e. it was nowhere close to what my numbers were -- counting 50% of transactions whereas I expected 90% of customers to see the page I was counting at). That has since cleared up. Fear of autopilot: They have pretty comprehensive budgeting control (you set it in terms of a per-day thing, but that is basically a marketing-friendly wrapper around a per-month budget). That lets you sharply limit the risk -- I think at the start I had a daily budget of $10, which means if the process went out of control for a month I'd be out at most $300 when I got the next bill and terminated it. They've since performed well enough that I've adjusted my budget upwards several times until they were unable to spend most of it. (The default for AdWords is to adjust your bidding to spend up to your budget -- for example, if you have $500 unspent, they'll adjust your bids upwards to get you more clicks and spend that last $500. This default is in Google's interests, but not in yours, because it will result in you paying absolutely stupid per-conversion costs at the end.) For this Halloween I'd feel comfortable authorizing them to spend several thousand dollars. (Halloween is my best season and its the only one where my ads ever get budget-constrained.) Ways it could have been easier: Conversion Optimizer is about as easy as I could want it to be. AdWords, hmm... I'd like there to be a clone of me who I can get to write creatives (that is advertiser-speak for ads and landing pages) and check if they're performing well without me having to actually do it. Everything about quality score could be more transparent, particularly for newbies to the system. Changes to improve ROI: I've always got ideas for this, I just don't have nearly enough time to implement and rigorously test them all. My big project right now is turning my desktop application into a web app, which I think will result in a substantial across-the-board increase in conversions due to the elimination of the download, install, execute cycle and the possibility to collect email addresses to implement an autoresponder & etc.
  - gfodor 17y ago
    thanks for the reply -- do me a favor, if you have time, I'd like to chat more offline, email me at gfodor at gmail if you're up for it!