On average, Finch clients reduce their cost per conversion by 58% and increase the volume of conversions by 67%. Every client is different, and Finch works with each client to establish priorities and target goals.
In most Google AdWords accounts, there is a portion of the keywords that are strong performers. Finch identifies those keywords and allocates more budget resources towards their increased performance. This provides the volume gain. At the same time, Finch identifies the keywords that have used a disproportionate amount of the budget for limited performance, and less budget is allocated to those keywords thus providing cost savings.
These achievements are further amplified by the Finch campaign structure that takes advantage of Google settings and choices, such as match types.
The magic box component is the Finch patent pending algorithms and data models that decide what to do when you don’t have perfect data. This is a key factor since even mature campaigns may only have less than 10% of the required data to be fully optimized.
Several million, but in general we recommend filtering out keywords that do not have any measurable activity.
Finch tends to max out the number of Ad Groups (with the standard limitation of 2k from Google), but we have accounts with over 100k split up in multiple campaigns. The restriction is Google and what Google allows. Finch runs in the cloud and has yet to hit a limitation in size of campaigns. (We have opted out of client engagement because of the relationship between volume and CPC due to Google API fees.)
Finch guards this information, but we run campaigns that are close to $1M per month. Finch has a range of clients from small (that spend from $5k per month) to global brands in our client base.
Setting CPA goals per campaign is counter-productive because of the non-linear relationship between keywords and products (actions taken) on the web site.
CPA goals (or better yet, CPV or CPP, which should be used for eCommerce companies willing to pay more for clicks returning more profit/revenues) must be set for each instance of the keyword and adjusted as new data comes in, because the competitive landscape tends to change frequently not only because of seasonality, but for other reasons as well (competitors bids, their quality scores, bidding rules they apply, their prices and terms, etc.), and to stay competitive while controlling costs, you must be very adaptive.
No, we cannot make estimates, because every click comes from a real time auction with a lot of other competitors. Nobody will ever now the impact unless we try changing the target.
Following are specific reasons why we cannot make estimates for future revenue:
The short answer is no. Finch does not look at assisted clicks (assist keywords). We only look at the last keyword clicked since that keyword is the one that converted.
It is not common for a path of clicks leading up to a conversion to be repeated, so if we allocate ad spend to an assist keyword, we would be subsidizing non-performing keywords and work against our optimization.
If we increase the CPC for an assist keyword, we would have to take away from a performing keyword since budget is likely a scare resource in any campaign.
In the case where budget is not a scare resource (in other words you always have more than enough budget to use on ad spend), Finch suggests that you increase the target per conversion (or CPV) to drive more conversions from existing performing keywords instead of allocating budget to non-performing assist keywords (measured by conversions).
Read more on this topic in the Finch blog article, 4 reasons why the PPC management click funnel strategy is a lie.
Yes, but you do not know which of your keywords fall into that category – whether or not you use Finch. The only viable strategy is to bid more on everything and hope that the potential customers come to the store. If you cannot track it, you cannot act on it or optimize it.
Finch runs in the cloud and uses Amazon EC2/S3 as our infrastructure, ensuring unlimited scalability and benefiting from Amazons security and performance abilities.