When measurement systems are based on cookies, it’s easy to overstate the reach and understate the frequency of your campaigns, as you might show your ad to a same individual three times if they use three different devices to browse the web. Traditional measurement solutions would count these visits as coming from three different people, which doesn’t give you an accurate view of your marketing results and campaign’s impact as regular tools track only clicks, which leads to discrepancies when ad performance is compared between Facebook’s analytics and regular web-analytics. Ideally, configure your web-analytics to use influenced conversions instead of only relying only on last-click based analytics, which is the default configuration in most analytics tools and only reveals which users made the desired action immediately, but not possibly on a different device some hours or days later.
Facebook considers an ad impression leading to a visit similarly to a click on an actual ad, which can be tracked by web analytics. This means that regular web-analytics is not fully able to track Facebook ad reach as they are able to track only clicks on ads. This might come as a surprise to marketers that are used to track campaign performance using a web-analytics tool such as Google’s as addressed by Facebook in this article and one of the biggest advertising technologies for Facebook ads, Smartly in this blog – both recommended and valuable reads thoroughly covering the topic of display ads.
View Attribution on Facebook
An ad impression is a view. When your ad loads and displays in front of a user on Facebook, that is one impression. If the same Facebook user visits your site during the next 24 hours after seeing an ad, this is considered as a visit similarly to a user who clicked an ad. Your regular web analytics will miss this visit as it is technically able to track only clicks on ads whereas Facebook is capable to track impressions resulting as a visit to the site also across different channels, mobile, tablet and desktop browser.
Nosto uses the minimum time period available for calculating the traffic attribution, which is currently 24 hours. This means that if a customer sees an ad 08:00am on 1st of January and returns the store 08:01 on 2nd of the same month and buys something, it is not considered as a sale through Nosto’s Facebook ads, but you can still track this from Facebook’s own analytics. Using the same example, if a customer sees an ad 08:00am on a mobile device while heading to work, and shops later the same day at 09:00pm using his laptop instead, that is treated as a Nosto influenced sale.
Web Analytics Click Attribution
Regular web-analytics tool such as Google’s is actually quite pessimistic with the visit and sales attribution because the user has to by default buy directly after clicking the ad, unless a longer conversion attribution window is used or more advanced tool, referring to influenced conversions is used. Web-analytics also loses the cross device conversions from Facebook because sales is not attributed correctly if the user for example clicks the ad on mobile and then later buys with laptop within the 24 hour sales attribution period. Facebook is able to track both events, meaning it’s normal to see more conversions on Facebook/Nosto than in regular web analytics. Facebook’s longer time period of 28 days which is in their analytics used by default, is on the other is quite optimistic. However, this is not used to track Nosto influenced sales as our attribution window is shorter: 24hours. Nosto’s 24h attribution window actually gives a good middleground between regular web-analytics pessimistic and Facebook’s optimistic view. A good analogue is that Nosto tracks orders that happen the same day, minimizing a chance for user accidentally returning to the site shopping.
Facebook Ad Analytics
As a Nosto Facebook ad user, you can also access Facebook’s own analytics on Facebook. Nosto relays the same statistics from Facebook to Nosto, meaning that you can use either analytics tool to track your ad campaign performance in more detail. In case you see even more conversions on Facebook than on Nosto’s web-admin, it is likely that your Facebook analytics uses longer sales attribution period than Nosto, for example the default 28 days, but since Nosto’s analytics uses 24 hour sales attribution by default, orders taking place after 24h period since a customer has seen an ad are not treated as a Nosto sale. For comparison purposes, you can move the attribution model to Facebook default by selecting 28day view.
In case you prefer using Facebook’s own analytics, also adjust the analytics view accordingly to use Purchase (Facebook Pixel) to track actual sales and total conversion value, which are not visible in the default analytics view. Read this article to learn how.
Tracking Orders Through Facebook Ads
Unfortunately Facebook doesn't support surfacing information about an individual order in the statistics consequently meaning than an individual order can't be tied to a campaign and an ad. You can track what products were sold through a campaign an ad, but user-level information is inaccessible hence it can't be reported in the analytics. Ad clicks leading to purchase conversions are by default tracked by web-analytics, meaning that these interactions leading to sales can be tracked using a web-analytics tool. However, view-attribution based purchase conversions are only tracked by Facebook.
How to Exclude Facebook Bot Traffic
When you launch new campaigns or new products are added to your product catalogue, Facebook reviews product page links using a bot called Facebot. This’ll show up in your web-analytics as traffic from United States, which you might not even target with your ads, unless you exclude the traffic separately.
Here’s a quick guide how to filter Facebook’s bot traffic:
1) In Google Analytics, click Admin on the top navigation bar
2) Select your View and go to Filters
3) Click Add filter button
4) Select Create new filter give the filter a name, such as “Filter Facebook Bot traffic”
5) Select Custom as Filter type
6) Define the filter rule as Exclude → Filter Field ** ISP Organization** → Filter patternfacebook inc. and leave Case-sensitive unchecked
7) Save the filter
8) Repeat steps 2 → 7 for any additional Views, if necessary