How Smart Are Smart Goals?

There’s an interesting debate on the newly introduced Smart Goals going on. Although this is the opposite of “Advanced AdWords”, I’d like to add a few points that haven’t been made yet.

Are Smart Goals dumb?

Julie Bacchini startet the discussion with a clear opinion: “Let’s Use Made Up Conversions“. Larry Kim, who’s company deals with many smaller advertisers, disagreed and called Smart Goals “Pretty cool for certain advertisers“. Not amused, Julie wrote a reply and got support from Bryant Garvin.

Basically, the discussion so far is to what extend Smart Goals work towards predicting the actual thing and whether that’s helpful or not. I wrote a post on our German blog today and looked at some data that I found pretty convincing.


menuWe found Smart Goals were already available in some of our client’s Analytics accounts. With those, Google has released a new report in Analytics. If your account is eligible you can find this report among your conversion goals reports.

The report provides only a few statistics but includes regular conversions. Looking at a client with Analytics conversions fully set up lets us evaluate the new Smart Goals pretty easily.

The full report looks like this:


As you can see, the stuff about behavior makes sense: Sessions with completed Smart Goals were much longer and had much more page views. But the conversion columns are much more interesting (I cut the rest out of the picture):


There are a few things to see here:

  • Overall conversion rate is 2.71% compared to 1.68% of sessions with a completed Smart Goal. The number of counted Smart Goal completions is about 40% lower than the actual conversion number.
  • Sessions with a completed Smart Goal actually had an ecommerce conversion rate of 34%. In other words, 66% of the reported Smart Goal completions are not Conversions at all.
  • Smart Goals only cover 21% of Conversions but 35% of revenue. This indicated that those longer, more intense sessions resulted in more expensive transactions. Makes sense.

Is this dumb or not?

In this case I find the data very convincing: Smart Goals are way off. Now an online retailer like this is certainly not representative of the kind of small businesses that are supposed to use Smart Goals, but it makes me wonder how it’s supposed to work any better for them. Maybe Larry Kim or Wordstream can confirm or refute this for smaller clients.

Possible design flaws

This example aside, I wonder if Smart Goals could even theoretically work for small businesses. When I imagine the kind of business that isn’t able to implement proper conversion tracking I envision the kind of business that you’d call or come by. However, both calls and store visits are heavily underrepresented conversion types in Google Analytics. Even though Google has tons of anonymized conversion data, I think the online retailer from the example above should be much better covered than the many small businesses from the offline world.

There’s anonther thing that I found strange. According to the Analytics help page, each visit to a website gets a score. Then a threshold is established by selecting the (approximately) top 5% visits from AdWords. This suggests that Googles assumes a 5% conversion rate from AdWords for every individual advertiser. On the other hand I checked several examples and found only about 2% of the AdWords sessions with a completed Smart Goal. So far I couldn’t make sense of this.

Less work for advertisers?

Google has Smart Goals intended for advertisers who don’t want to deal with setting up conversion tracking. This is also the point that Larry Kim makes: For advertisers who wouldn’t set up conversion tracking this is a much easier alternative. I’m not so sure about that.

Using Smart Goals requires a few steps:

  • Link Analytics and AdWords
  • Enable data sharing in Analytics
  • Enable Smart Goals
  • Import them into AdWords
  • Use them in AdWords, possibly even for automated bid strategies

I admit I’m probably not the right person to judge what is easy and what is not when it comes to stuff like this, but it seems to me that setting up conversions in Google Analytics isn’t that much harder compared to Smart Goals. Both of them require someone to learn a bit about what they’re doing and then click through more settings that they’re comfortable with.

My advice for small businesses

Even if you don’t want to put too much effort into learning about conversion tracking and web analytics and everything, don’t put up with Smart Goals. If you’re an AdWords Customer, get Google Analytics on your site and then call the AdWords support. Have them walk you through a proper setup for Conversions in Analytics. I haven’t tried this specifically, but in my experience they’re very helpful with things like this.

About Martin Roettgerding
Martin Roettgerding is the head of SEM at SEO/SEM agency Bloofusion Germany. On Twitter he goes by the name @bloomarty, and you can find him regularly on #ppcchat.

  • Jordan McClements

    I’ve very little data to work with here – but from my own accounts so far it *seems* that smart goals are nothing more than a goal which says “traffic with a *very* low bounce rate and high number of page views”.

    I guess it saves you 5 minutes setting up a ‘non bounce’ goal in Analytics…? (A goal which I have tried but haven’t found to be a huge help).

    Your data is lot more flattering for smart goals than mine.

    So far, for a lead gen account with a really good conversion rate and lots of traffic the ‘Yes’ smart goal traffic accounts for only 2.39% of conversions! So to optimise for this traffic only would be almost completely pointless. Unless I am missing something?

    • Martin Röttgerding

      Hey Jordan,
      Thanks for sharing this. So far my data is indeed more flattering…

      I think your definiton fits very well. Basically Smart Goals seem to be the most intense sessions aided by a few other less important signals. Maybe an iPhone user’s session can be a little shorter than an Android session to count – something like that.

  • Larry Kim

    HI. thanks for doing this. as i mentioned on Twitter, i haven’t even had a chance to test it out yet! here’s what i know. 2 years ago, my product development team and I were looking into coming up with a way of developing bid management tools that worked without having conversion goals installed, which is a real challenge for half of our prospects and even a good number of customers. We looked at quite a few accounts (a few thousand) and there was indeed a relationship between engagement rates (time on site, bounce rates, pages per session, etc.) and conversions. though, it’s like Quality Score – directionally aligned across a large number of accounts, but individual results vary.

    Even with your example, it’s interesting that a small number of smart goal completions converted at a very high rate. i realize it was a small percentage of total number of conversions, but sessions with smart goal completions in your data set appear far more likely to convert to real conversions. Thus, if a company was limited by budget, it would make sense to prioritize the smart goal completions as the “low hanging fruit”.

    Finally, i’d note that you’re comparing regular conversion tracking vs. smart conversion tracking. I’m not surprised that the real thing is better than the estimated thing. but it’s possible that smart conversions might be better than nothing at all.

    anyway, thanks for writing it up. good work.

    • Martin Röttgerding

      Hi Larry,
      Thanks for commenting! This bid management challenge sounds fascinating. I wonder if this could be improved by looking at individual business types… maybe there are commonalities, like all doctor’s offices can rely on the same set of signals. Then again I suppose it depends on websites as well. A website with important information like address and phone number right on the home page may have totally different signals than a site that buries that info somewhere. Intriguing challenge.

      Hm, I don’t know if I agree with your point about going after Smart Goals if the budget is limited. It seems to make a lot of sense, but I can’t quite wrap my head around it. If you combine bid management with Smart Goals, aren’t you basically looking for signals that correlate with signals that correlate with conversions? Maybe I’m thinking too complicated here…

      I also wonder if there isn’t a way to determine if Smart Goals are better than nothing or not. Guess I’ll have some thinking to do (it should be about Christmas presents, but this is way more interesting).

  • James Keeney-wilson

    Interesting article, I’m in the same boat of struggling to wrap my head around this. I’ve been looking through a few accounts and have mixed data, some the smart goal data is flattering others not so.

    I think there were 2 ideas behind it. The first being that if more of your AdWords traffic were what i’m going to call smart traffic (users that had completed a smart goal). Then the overall conversion rate would be improved. Compared with your current optimisation AdWords method.
    So while in Jordans example smart goals may (at the moment) only account for 2% of all conversions that traffic may have a conversion rate of 33% instead of 2% for all site traffic. So if all the AdWords traffic was this smart traffic then in theory the conversion rate would be as high 33%.

    The Second idea behind it, i think, is that there are a lot of small business out there that are simply unable to set up proper conversion tracking because they don’t have contact forms or developers haven’t created thank you pages for forms, or installed event tracking on email links. That or the businesses are unwilling to install on-site call tracking, so for these users implementing smart goals is easier than having a developer change the site.

    I’ve included a screen of some smart goal data from a low converting ecommerce site, as you can see the conversion rate of the paid smart traffic is significantly higher than the conversion rate of the non-smart paid traffic. So more of that traffic would be ideal.

    I still don’t know how useful smart goals are going to be and i will be thinking on it over Xmas too, but more data is always useful.

    • Martin Röttgerding

      Hi Kevin,
      Thanks for your comment! I think that only having this “smart traffic” sounds good, but it can’t be done. This is what I had to wrap my head around after Larry’s comment.

      Smart Goals are treated like any other kind of conversion in AdWords. For example, if 5% of your AdWords traffic results in a Smart Goal completion that means you see a 5% conversion rate in your AdWords account. Only wanting smart traffic would mean you’d only want the 5% of clicks that convert. Smart Goals or regular conversions, this is basically what we all want 🙂

  • Larry Kim