Changes to ECPC
By removing the 30% bid cap on ecpc bids, what can I do to prevent clicks at extremely high bid amounts set by other advertisers? Example:
Let's assume I'm running a shopping campaign and am bidding .30 for a product while advertiser A has bid set to .90 for same item. Previously, my bid would've competed with other advertisers and "beat" those bidding under .39 (theoretically). Even if that .30 bid doesn't win every auction, at least I'm not paying more on bids that may not convert.
Just wondering if I need to worry about reducing my existing bids across the board.
Re: Changes to ECPC[ Edited ]
May 2017 - last edited May 2017
the intent of removing the cap on the ecpc algorithms is mainly
to accommodate any outliers related to audiences and locations --
while respecting the manual-bid and maintaining cost-per-conversion,
or in the case of shopping-ads, maintaining the overall spend amount.
the audiences and locations are key --
previously, bid-adjustments were needed
since those targets were not considered;
mobile-devices are still not considered.
for most shopping-ad use-cases, ecpc is less than ideal except for merchants
who have large inventories or require large numbers of manual bid adjustments,
with enough performance history and related data for the algorithms to behave
optimally -- the machine-learning algorithms tend to favor large data-sets over
long time intervals.
since the manual-caps are still respected, there should
be little change to direct bid-to-bid related competition.
however, what is more likely to change are audience and location competition --
for example, if a shopping-ad had little competition from a particular locale,
there may be an increase in overall bid competition where little existed before.
conversion-rates are another critical component to ecpc --
if competitors have drastically different conversion-rates
then the algorithms are less likely to have an impact.
generally, keep track of performance measurements with respect to
conversion-rates, remarketing, audiences, and locales; then, decide
if countermeasures may be needed based on those measurements.