Here's my log:
Request:
...
<SOAP-ENV:Body>
<ns2:get>
<ns2:selector>
<ns2:searchParameters
xsi:type="ns2:RelatedToKeywordSearchParameter">
<ns2:keywords>
<ns1:text>papo knights</ns1:text>
<ns1:matchType>BROAD</ns1:matchType>
</ns2:keywords>
</ns2:searchParameters>
<ns2:ideaType>KEYWORD</ns2:ideaType>
<ns2:requestType>IDEAS</ns2:requestType>
<ns2:requestedAttributeTypes>KEYWORD</
ns2:requestedAttributeTypes>
<ns2:paging>
<ns1:startIndex>0</ns1:startIndex>
<ns1:numberResults>20</ns1:numberResults>
</ns2:paging>
</ns2:selector>
</ns2:get>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>
and here's the response:
...
<soap:Body>
<ns2:getResponse xmlns="https://adwords.google.com/api/adwords/cm/
v200909" xmlns:ns2="https://adwords.google.com/api/adwords/o/v200909">
<ns2:rval>
<ns2:totalNumEntries>477</ns2:totalNumEntries>
<ns2:entries>
<ns2:data>
<ns2:key>KEYWORD</ns2:key>
<ns2:value xmlns:xsi="http://www.w3.org/2001/XMLSchema-
instance" xsi:type="ns2:KeywordAttribute">
<ns2:Attribute.Type>KeywordAttribute</
ns2:Attribute.Type>
<ns2:value>
<Criterion.Type>Keyword</Criterion.Type>
<text>papo figures</text>
<matchType>PHRASE</matchType>
</ns2:value>
</ns2:value>
</ns2:data>
</ns2:entries>
<ns2:entries>
<ns2:data>
<ns2:key>KEYWORD</ns2:key>
<ns2:value xmlns:xsi="http://www.w3.org/2001/XMLSchema-
instance" xsi:type="ns2:KeywordAttribute">
<ns2:Attribute.Type>KeywordAttribute</
ns2:Attribute.Type>
<ns2:value>
<Criterion.Type>Keyword</Criterion.Type>
<text>papo figures</text>
<matchType>BROAD</matchType>
</ns2:value>
</ns2:value>
</ns2:data>
</ns2:entries>
any ideas?
The match type specified in the RelatedToKeywordsSearchParameter only
influences the way that keywords ideas are generated, but it doesn't
restrict the match type of the returned ideas. If you only want to
have certain match types returned you must use the
KeywordMatchTypeSearchParameter as well:
Best,
- Eric Koleda, AdWords API Team
I have two main questions based on your post:
> The match type specified in the RelatedToKeywordsSearchParameter only
> influences the way thatkeywordsideas are generated,
1) What do you mean? Say I have the type set to BROAD - only keywords
with BROAD relation to the searched keyword are returned?
>but it doesn't
> restrict the match type of the returned ideas. If you only want to
> have certain match types returned you must use the
> KeywordMatchTypeSearchParameter as well:
2) Why infact - how are the results to be returned determined? As
described in 1) I have only searched for BROAD related keywords. What
does it mean now if a generated keyword has the type EXACT? Is the
relation BROAD or EXACT?
Example:
Selector: cat RelatedToKeywordsSearchParameter type = BROAD
Result: cat something KeywordMatchType = EXACT or BROAD or PHRASE
So what is the difference in the realtion when asking:
Selector: cat RelatedToKeywordsSearchParameter type = BROAD
cat KeywordMatchTypeSearchParameter type = EXACT
Result: cat something KeywordMatchType = EXACT
I would be happy to get some infos on this topic.
Thanks in advance,
Dian
P.S. I got in this issue by checking why do I get duplicate results on
keyword generation [http://groups.google.de/group/adwords-api/
browse_thread/thread/
dfb3c8928390c553/3af5711a0efe93a9#3af5711a0efe93a9]
I understand that this functionality is confusing, and I'll do my best
to clear it up.
1) The TargetingIdeaService uses a complicated backend to determine
relevant keywords, and I don't have the full information about how it
works. I'm not sure how much weight, if any, is put into the match
type of the source keyword. To be clear though, this match type may
influence the results generated by the TIS, but it doesn't guarantee
that all the results are of that match type.
2) Keywords can come in any of the three match types, and they all may
be related to the source keyword. This is why you can get an EXACT
keyword idea as a results for a BROAD source keyword.
Best,
- Eric
thank you for your reply but I need a more precise explaination of
this functionality in order to get the best results for generating
campaigns - the lack of knowledge on this topic is not an option for
me.
I did the following test:
* I picked a keyword and tested 3 combinations
** source keyword from type EXACT, results from type BROAD (eb)
** source keyword from type BROAD, results from type EXACT (be)
** source keyword from type BROAD, results from type BROAD (bb)
* I generated a couple of results (see below), compairing them based
on AVERAGE_TARGETED_MONTHLY_SEARCHES (AM), GLOBAL_MONTHLY_SEARCHES
(GM) and the Rank (based on the order in which the results are
returned):
Text ebAM beAM bbAM ebGM beGM bbGM ebRank beRank bbRank
kw1 33100 14800 33100 40500 14800 40500 0 0 0
kw2 390 0 390 390 91 390 5
1 1
kw3 880 210 880 880 210 880 1 2 2
kw4 2900 480 2900 4400 480 4400 2
3 3
....
kwx 74000 22200 74000 74000 27100 74000 682
682 790
kwy 8100 590 8100 8100 590 8100 683 683 792
I just can not understand how are the results generated. As you can
see:
eb = bb regarding volume
be = bb regarding rank (sometimes)
eb = be regarding rank (sometimes)
Could you (or some one from the developers) explain shortly based on
examples why the difference in the results. I would like to see at
least how are the dependencies between the match type and the rank /
search volume / competition / CPC defined.
Best regards,
Dian
As I mentioned before, the match type of the source keyword doesn't
affect the match type, and therefore statistics, of the returned
keywords. If you ignore the match type of the source keyword from
your tests you can see the statistics only change based on the match
type of the returned keyword.
Best,
- Eric