Yahoo! 'Assists' Their Advertisers On PPC Bidding
First let me break down what Assists are. Assists are simply keywords that played a part in assisting other keywords in getting a conversion.
Huh?
Ok, here's an example:
Let's say I'm a visitor looking to purchase a computer. I go to Yahoo.com and run a search using the term "computer" . I decide on clicking the Sony sponsored search ad for computers. I dig around on the SonyStyle.com web site but end up not purchasing anything. I then come back 14 days later and use the search term "Laptop" on Yahoo.com. But this time I decide to click on a different Sony sponsored search ad....for laptops. I reach the same SonyStyle.com site and I finally decide to make a puchase (conversion).
The results in Yahoo....
The term "computer" will be counted towards an Assist
The term "Laptop" will be counted towards a Conversion
Notes:
1. Assists are tied to conversions for up to 45 days. If my first search was "computer", that term will be counted as an Assist as I long as I convert within 45 days following that first search.
2. The visitor must click through one or more of the same advertiser's ads before the Assist will be tied to the Conversion. For the above example, the visitor clicked on the Sony ad for 'computers' and then came back and clicked on the Sony ad for 'laptops'.
3. The advertiser must be using Yahoo's 'Conversion Only' analytics tag or Yahoo's 'Full Analytics' tags on their confirmation/thank-you page.
Great description Matt, but how do I make use of Assists for bidding on my keywords??
Let's first take a look at what the Assists report looks like in Yahoo...
When looking at the keyword level in your Yahoo reports (in this case it's called 'Marketing Activity') you see that the keywords are broken down into Leads (# of times visitors reached your landing page), Assists, and Conversions (# of visitors who reached your confirmation/thank-you page).
With traditional analytics reporting, advertisers will typically look at the keyword to see how many conversions it received to determine how they want to bid on that keyword. In this case, they would see that "Raiders Tickets" had received zero conversions (I am a Raiders fan, so this was a tough example for me to cough up!). With no conversions, typically the advertiser would either lower the bid on the keyword or completely dump the keyword as a "non-performer".
However, having an Assists column adds a whole new dimension to the way we traditionally think about bidding on terms. You'll notice that while the term "Raiders Tickets" didn't convert, it did drive 33% of the Assists. This means the term "Raiders Tickets" played a part in assisting other keywords in getting a conversion.
Think about it in basketball terms. A point guard might get 20 assists but no points (conversions) in one night. That's still a pretty productive night for the point guard! Are you going to bench your point guard after he dished out 20 assists and helped the other players score 40+ points?!?! No way! That point guard is valuable to the team.
Same theory goes into thinking about your keywords. Before you lower the bid on a particular keyword you should ask yourself, "Just because this keyword didn't convert well for me, did it help other kewords convert? If so, do I really want to lower the bid on a keyword that drives a high percentage of my total Assists?" Probably not. More likely, you'll want to continue to bid on this keyword as it's helping drive conversions and brand awareness.
So really, Assists help out on two fronts:
1. They Assist other keywords in converting and
2. They Assist YOU, the advertiser, in making smarter bidding decisions.
You can also find a great post on Yahoo's Assists at: http://komarketing.typepad.com/blog/2007/01/yahoos_conversi.html
Comments
So where does this information relate to? Where does it come from? Thanks Sebastian
You shouldn't receive any data starting the 46th day from when you turned the campaigns off.
I have a question, normally the assist counts are bigger than conversion, why is this happening?
cause I bet assist would be counted in a case coversion happens, so the counts of conversion and assist would be same or conversion could be bigger.
Good question. Sometimes you might see 2 Assists for a campaign and 0 Conversions for example and wonder, "How can this be?"
This happens because the Assists are being attributed to other campaigns and not necessarily the campaign you are looking at.
For example, let's say I'm looking at the data for Campaign A. Over the past two days I see that it has received 0 Conversions (meaning visitors probably clicked on the ad for Campaign A but never bought anything). However, Campaign A does show 2 Assists. This means that Campaign A was used in "assisting" one or more of your other campaigns in a conversion (it could have helped campaign B convert twice for example or maybe it helped covert campaign B once and also campaign C once).
In the next version of the Assist report we'll be adding a deeper layer of Assist detail so you be able to see exactly what campaigns the Assists attributed to. For example, let's say you see Campaign A had a total of 4 Assists. We plan to show you that of those 4 Assists, they resulted in one conversion for Campaign A (because a campaign can Assists itself), two conversions for Campaign B, and one conversion for Campaign C. Totaling 4 "Assisted" Conversions.
For ecommerce sites, we're looking to add value information for Assists also. For example, let's say that we find that Campaign A again had 4 Assists. Of the 4 Assists we learn that 2 of those Assists went to helping convert Campaign B and the other 2 Assists went to helping convert Campaign C.
While this deeper level of information is good, how do we know how valuable those Assists really are in terms of driving revenue to other campaigns? Campaign B and C both received two Assists each from Campaign A. So does that mean that should be treated as equal value? No.
By passing us revenue information in the Yahoo Analytics conversion tag, we'll be able to show you the revenue output of Assists. For example, we'll be able to show you that while Campaign B and C both received two assisted conversions from Campaign A, Campaign B had a revenue total of $100 while the two Assists that went to Campaign C only generated $50.
Therefore Campaign A assisted in driving more revenue for Campaign B than it did for Campaign C.
It's information that nobody has had before when trying to optimize their campaigns. Just imagine not having the Assist report and dropping the bid on a keyword that had low conversions only to realize later that the keyword had 25 Assists associated with it and it ended up driving $2500 in assisted revenue.
You'd probably bang your head against the wall a few times!
Hope this helps.
great article thanks,
I have a question re Assists ? how do I measure 1st,2nd,3rd etc assists in google addwords, is it a standard report ?
kind regards
subhas
You'll want to look into Google's Conversion Funnels to get Assists.
Matt