Sunday 27th May 2018

Facebook Track: Top 100

Summary

This page is intended to act as a supplement to the “Facebook Track: Top 100” report which is available to download here (PDF 3.6MB). This page explains the methodology in more detail and provides some additional data to compliment the data in the report.

Methodology

Facebook offers brands a number of options when it comes to creating pages. 60 percent of the top 100 brands analyzed utilize a Facebook feature called “global pages”. Global pages permit a brand to maintain a “default” page e.g. https://www.facebook.com/facebook and “market pages” which are targeted to specific countries or regions e.g. https://www.facebook.com/facebookindia

For brands which have implemented global pages, a user can switch between the “default page” and a “market page” by selecting “switch region” on a Facebook page.

The default page is often targeted to a small subset of countries and the brand will create market pages for all other countries or regions it wishes to target. This poses a problem for data analysis because the only metric made available by Facebook is “page likes” and this metric aggregates all likes from all default and market pages for a brand i.e. it tells us nothing about the number of likes for a particular page.

Moreover, if we were to analyze engagement rate for a default page e.g. /netflix/, then we would be analyzing the engagement rate for Netflix’s entire audience for posts which only went to a small subset of that audience (the countries its default page is targeted to).

For example, Netflix, at the time of writing, has roughly 44,000,000 fans and it has made over 500 posts in the last year. The problem is, those posts aren’t seen by Netflix’s 44,000,000 fans, they are only seen by the countries being targeted by that default page.

To add further complication, some brands have implemented regional pages independently of Facebook’s global pages structure. These pages are less of a concern as likes to these pages aren’t aggregated alongside all other regions. In instances such as these, we’ve typically elected to analyze a brand’s main, non-regional page.

How do you calculate page likes of a single market page?

We used online tools to determine how many likes a market page received from a particular location. For example, we determined that the Netflix US page received 9,275,147 likes from the U.S. Therefore, when it comes to calculating engagement rate for Netflix, we selected the Netflix US page and then divided total engagements by number of US page likes.

In a small number of cases, and where a brand had relatively few market pages, we simply analyzed the number of likes from the countries targeted by its market pages and subtracted this total from total page likes (all of the remaining countries were targeted by the default page).

For example, Gucci only has two market pages: South Korea and Japan. We looked at the audience share of likes for these two countries and subtracted them from Gucci’s total page likes. We then pulled all posts from Gucci’s default page between March 15th 2017 and March 15th 2018 and calculated total engagement based on Gucci’s new page likes total (minus page likes from South Korea and Japan).

What do you classify as an engagement?

An engagement in this analysis is defined as a comment, like, reaction or share.

What does the “pages and global/market pages” category contain?

You can view the full list of pages in this category below. Where a brand hasn’t implemented Facebook’s global pages structure, we’ve obviously analyzed the same page between both categories. Where a brand has implemented Facebook’s global pages structure, we’ve selected a market page to be representative of that brand.

The “pages” category, for the purposes of this report, has really only been utilized to compare total page likes between brands on a global level i.e. aggregate page likes of all brands.

In all instances we’ve endeavored to select active Facebook pages from countries with large populations.

The table below contains 100 pages, 60 of which are market pages or default pages and have been selected to represent a brand. Apple has been excluded from the “pages and global/market pages” category as it doesn’t maintain an active Facebook presence.

Brands which have implemented Facebook’s global pages structure are emboldened.

  • Hewlett Packard Enterprise has implemented global pages however it has only made a default page available i.e. it doesn’t provide additional market pages. For this reason we chose to analyze its default page and its global aggregated page like total, despite the fact it appears in the “pages and global/market pages” category below.
  • Loreal Paris has titled its default page “LorealParisUSA” (despite the fact it targets multiple countries). Similarly, Samsung’s default page is titled “SamsungUK” (/samsung/ redirects to /samsungelectronics/, a page with significantly fewer likes than /samsunguk/)
  • We weren’t able to pull audience share data for Zara, as it has chosen to use a page ID in its URL. However, it only uses two market pages so we selected the default page for the “Pages and global/market pages” category.
  • We had trouble pulling reliable data on Discovery from Facebook’s API. We ended up combining data from two sources and filtering duplicates.
  • It may appear as though non-intuitive market page selections have been made for a few brands e.g. we selected FedEx Canada rather than FedEx US, despite FedEx being a US corporation. This is largely a result of brands electing to select their home country as a default page and targeting upwards of hundreds of other countries from this default page. For example, if we had selected FedEx US — the US is a target country for FedEx’s default page — we would need to have analyzed audience share for the upwards of hundreds of other countries that default page targets.
  • We selected the UK and Ireland market page for Pampers and included page like totals from both countries.
  • We included Tesla in our analysis despite the fact it recently deleted its Facebook page. Unfortunately, we had only pulled its post likes prior to its deletion, therefore its engagement rate (0.18 percent) — see the report — is only based on post likes, not all post interactions (likes, comments, reactions and shares).
  • In instances where a brand maintains a corporate page and a consumer page i.e. Thomson Reuters maintains /thomsonreuters/ and /reuters/, Huawei maintains /huawei/ and /huaweimobile/ and JP Morgan maintains /jpmorgan/ and /jpmorganchase/, we’ve typically opted to select the more active presence.
 Brand “Pages” “Pages and global/market pages”
Moët & Chandon /moet/ /moetusa/
Smirnoff /smirnoff/ /smirnoffus/
Johnnie Walker /JohnnieWalker/ /JohnnieWalkerUS/
Corona /Corona/ /coronaextrabrasil/
John Deere /JohnDeere/ /JohnDeereUSCA/
Shell /Shell/ /ShellBrasil/
Sprite US /Sprite/ /spriteus/
Mini /MINI/ /MINIUSA/
Burberry /Burberry/ /Burberry/
Heineken /heineken/ /HeinekenUK/
Salesforce /salesforce/ /SalesForceUK/
KFC /KFC/ /KFCBrasil/
Jack Daniels /jackdaniels/ /jackdanielsuk/
Tiffany & Co. /Tiffany/ /TiffanyAndCo.Brasil/
PayPal /PayPal/ /PayPalUSA/
Netflix /netflix/ /netflixus/
Land Rover /landrover/ /landroverusa/
FedEx /FedEx/ /FedExCanada/
Mastercard /Mastercard/ /MastercardUS/
Huawei Mobile /huaweimobile/ /HuaweimobileMX/
Lego /LEGO/ /LegoGermany/
Cartier /Cartier/ /cartier.usa/
Visa /Visa/ /VisaUnitedStates/
Nestle /Nestle/ /nestle.USA/
3M /3M/ /3MDeutschland/
Adobe /Adobe/ /AdobeIndia/
Adidas /adidas/ /adidasUK/
Danone /Danone/ /DanoneWaveNA/
Gucci /GUCCI/ /GUCCI/
Allianz /allianz/ /allianz/
HSBC /HSBC/ /HSBCUK/
Loreal Paris /LOrealParisUSA/ /LOrealParisUK/
Philips /Philips/ /PhilipsIndia/
Volkswagen /Volkswagen/ /VolkswagenUK/
Nissan /nissan/ /NissanUK/
Nescafe /Nescafe/ /NescafeUK/
Ford /ford/ /forduk/
Hermes Paris /hermes/ /hermes/
Budweiser /Budweiser/ /BudweiserUSA/
UPS /ups/ /ups.uk/
Pampers /PampersGlobal/ /PampersUKIre/
American Express /AmericanExpress/ /americanexpressmexico/
IKEA /IKEA/ /IKEAUSA/
Zara See above See above
HM /hm/ /hmunitedkingdom/
Pepsi /pepsi/ /PepsiUS/
SAP /SAP/ /SAPIndia/
Cisco /Cisco/ /CiscoUKI/
Intel /Intel/ /IntelUK/
Disney /Disney/ /DisneyUK/
McDonalds /McDonalds/ /McDonaldsUS/
General Electric /GE/ /GEIndia/
IBM /IBM/ /IBM/
Facebook /facebook/ /FacebookIndia/
SamsungUK /SamsungUK/ /SamsungIndia/
Coca-Cola /Coca-Cola/ /CocaColaUnitedStates/
Microsoft /Microsoft/ /MicrosoftIndia/
Google /Google/ /GoogleUK/
Kelloggs /KelloggsMexico/ /KelloggsMexico/
Hewlett Packard Enterprise /HewlettPackardEnterprise/ /HewlettPackardEnterprise/
Lenovo /lenovo/ /lenovo/
Tesla /tesla/ /tesla/
Dior /Dior/ /Dior/
Prada /Prada/ /Prada/
Caterpillar /caterpillar/ /caterpillar/
Ferrari /Ferrari/ /Ferrari/
Discovery /Discovery/ /Discovery/
Harley Davidson /harley-davidson/ /harley-davidson/
DHL /dhl/ /dhl/
Panasonic USA /Panasonic/ /Panasonic/
Johnson Johnson /jnj/ /jnj/
Kia /kia/ /kia/
Santander Brasil /santanderbrasil/ /santanderbrasil/
Reuters.com /Reuters/ /Reuters/
Morgan Stanley /morganstanley/ /morganstanley/
Colgate /Colgate/ /Colgate/
Sony Mobile /sonymobile/ /sonymobile/
Starbucks /Starbucks/ /Starbucks/
HP /HP/ /HP/
CanonUSA /CanonUSA/ /CanonUSA/
Siemens /Siemens/ /Siemens/
Porsche /porsche/ /porsche/
Citi /citi/ /citi/
Goldman Sachs /goldmansachs/ /goldmansachs/
AXA US /AXAUS/ /AXAUS/
Audi /audi/ /audi/
Accenture /accenture/ /accenture/
Hyundai Worldwide /Hyundaiworldwide/ /Hyundaiworldwide/
eBay /ebay/ /ebay/
JP Morgan Chase /jpmorganchase/ /jpmorganchase/
Gillette /gillette/ /gillette/
Honda /Honda/ /Honda/
Louis Vuitton /LouisVuitton/ /LouisVuitton/
Nike /nike/ /nike/
Oracle /Oracle/ /Oracle/
BMW /BMW/ /BMW/
Mercedes-Benz /MercedesBenz/ /MercedesBenz/
Toyota /toyota/ /toyota/
Amazon /Amazon/ /Amazon/
Apple /apple/ /apple/

Outliers

Listed below are the 74 posts (outliers) excluded from analysis. The posts are ordered by post engagement rate.

From the 29,047 posts analyzed it was clear a very small subset of outliers with hundreds of thousands of likes were skewing the data (there’s a histogram charting the distribution of likes in the main report).

With Facebook, a brand has the opportunity to promote or boost its posts. While such activity shouldn’t skew the average to any meaningful extent, we observed data points with a disproportionately large number of likes relative to shares.

When reporting on average engagement or interactions, some organizations simply remove the top five percent of posts. In our case, we segmented outliers that were three standard deviations beyond the mean and then created a new metric to analyze total post shares as a percentage of total post likes.

From this, we identified 74 posts belonging predominantly to one brand where shares were less than 0.5 percent of total post likes. We treated these posts as outliers.

The main report analyzes data both containing and excluding outliers and indicates where outliers have been excluded from analysis.

Page Post likes Page likes Engagement rate
citi 634,027 1,159,695 54.74%
citi 611,725 1,159,695 53.36%
citi 544,863 1,159,695 47.07%
citi 475,693 1,159,695 41.09%
philipsindia 414,745 1,123,895 37.09%
citi 341,518 1,159,695 29.48%
citi 326,360 1,159,695 28.17%
siemens 135,850 503,197 27.22%
citi 246,477 1,159,695 21.28%
siemens 101,267 503,197 20.25%
siemens 69,797 503,197 13.99%
siemens 68,899 503,197 13.78%
lenovo 834,420 6,704,037 12.59%
siemens 61,594 503,197 12.32%
allianz 117,266 1,066,346 11.01%
lenovo 688,348 6,704,037 10.40%
allianz 110,421 1,066,346 10.36%
allianz 103,343 1,066,346 9.70%
allianz 96,809 1,066,346 9.08%
allianz 90,155 1,066,346 8.46%
allianz 83,506 1,066,346 7.84%
lenovo 472,181 6,704,037 7.13%
allianz 68,419 1,066,346 6.47%
lenovo 430,262 6,704,037 6.47%
lenovo 416,689 6,704,037 6.26%
lenovo 398,883 6,704,037 6.01%
lenovo 370,183 6,704,037 5.56%
lenovo 298,594 6,704,037 4.50%
lenovo 298,425 6,704,037 4.48%
lenovo 285,565 6,704,037 4.28%
lenovo 279,517 6,704,037 4.20%
lenovo 268,591 6,704,037 4.06%
lenovo 268,503 6,704,037 4.03%
lenovo 262,728 6,704,037 3.96%
lenovo 260,850 6,704,037 3.93%
lenovo 246,450 6,704,037 3.70%
lenovo 221,547 6,704,037 3.31%
lenovo 213,591 6,704,037 3.20%
lenovo 212,685 6,704,037 3.19%
lenovo 207,059 6,704,037 3.13%
lenovo 198,982 6,704,037 2.99%
lenovo 191,703 6,704,037 2.89%
lenovo 192,085 6,704,037 2.88%
lenovo 178,110 6,704,037 2.69%
lenovo 174,324 6,704,037 2.62%
lenovo 166,320 6,704,037 2.49%
lenovo 156,875 6,704,037 2.35%
lenovo 155,163 6,704,037 2.33%
lenovo 149,896 6,704,037 2.25%
lenovo 144,037 6,704,037 2.17%
lenovo 139,496 6,704,037 2.10%
lenovo 138,656 6,704,037 2.08%
lenovo 137,697 6,704,037 2.06%
lenovo 132,712 6,704,037 1.99%
lenovo 130,432 6,704,037 1.96%
lenovo 124,056 6,704,037 1.87%
lenovo 124,329 6,704,037 1.86%
lenovo 118,126 6,704,037 1.78%
lenovo 115,453 6,704,037 1.73%
hyundaiworldwide 75,113 4,654,686 1.62%
lenovo 102,953 6,704,037 1.55%
lenovo 95,312 6,704,037 1.43%
lenovo 91,597 6,704,037 1.39%
lenovo 90,295 6,704,037 1.36%
hyundaiworldwide 60,984 4,654,686 1.33%
lenovo 85,466 6,704,037 1.29%
samsungindia 164,182 13,262,290 1.26%
lenovo 83,472 6,704,037 1.25%
lenovo 77,652 6,704,037 1.17%
lenovo 68,506 6,704,037 1.03%
lenovo 61,020 6,704,037 0.92%
samsungindia 117,011 13,262,290 0.89%
samsungindia 105,134 13,262,290 0.80%
samsungindia 82,247 13,262,290 0.63%