Mary, let’s get contrary with a slightly spurious multiplier
Inspired by a challenge from Dominic Mills, Route’s Euan MacKay outlines how Mary Meeker could create a size-based scale factor to improve her time/money analyses
In his column this week, Dominic Mills advocated including a scale factor in Mary Meeker’s time/money analyses. He has a point – it would take 53,000 iPhone 7 banner ads to fill a large billboard ad.
While admittedly a little crude, it did pique my inner geek…
Including a size-based factor to the calculation is an interesting one. Particularly with an Out of Home hat on, where size of canvas is virtually unparalleled across any other medium.
So, what did Dominic actually ask?
“Me, I’d like to add a plea that Meeker changes the basis of her calculation so that it factors in the relative impact of advertising in each as based on the size of the screen or available canvas (to account for print and OOH).
Thus, comparing the size of my mobile screen with that of my TV, Meeker should discount mobile by a factor of, say, 30 (or multiply TV by that figure); print, based on A4 for magazines and larger for newsbrands, should have discount/multiplier of between 6 and 20; tablets about 5; and OOH…well, I have no idea but you get the picture.”
Creating a Scale Factor for OOH
Surface area is vital in Route’s calculations for the audience measurement of Out of Home, so we’re uniquely positioned to help in response to this request.
We use the surface area to help determine:
- The maximum distance from which you can actually see a poster or screen
- The likelihood that it will be seen, when people are within the catchment areas
Route holds the surface area information for each of the 381,416 ads which it currently measures. Using these, and adding in a few other assumptions (which we’ll get to), we are able to complete Mills and Mackay’s Mary Meeker Mobile Multiplier (MMMMMM) for short.
These can be seen in the table below:
How does this work for OOH?
The average (mean) OOH ad size in GB has a surface area of 2.072 metres square. However, the inventory available to advertisers is highly varied. This includes small tube cards to large format ‘greater than 96 sheet’ digital Transvision screens and everything in between.
To give a flavour of the variance in size, the smallest poster measured on Route is 0.12m² and the largest single site is 600m².
Indeed, should you cover the palms of the good folk at Ocean Outdoor with sufficient quantities of gold, you could have yourself each of the iMax sites and create the largest out of home site in the UK which measures 1,734 m² (or c.2.7million square inches) in creative canvas, just as McDonald’s did recently. That’s the equivalent to 207,933 iphone 7s screens, should you be counting.
With such a highly varied dataset in Route, it perhaps makes more sense to even some of the skews out in our measure. For this we take the median (middle) size as a more appropriate metric to base our multiplier on. With this, the typical OOH site is 0.47m². The median digital out of home screen size is somewhat larger at 1.4 m².
It’s worth noting though, that this is still somewhat suppressed by the c.169,000 train/tram/tube cards which measure 0.13m² and bring this down. If these were excluded from the calculation, the average size of an OOH space in GB would measure 3.53 m² offering a mobile multiplier of 513.
Below are the mean sizes and multipliers for OOH inventory split out by environment and type (whether it’s a poster or a screen). You can see from this that there is significant variance by environment with some offering massive multipliers versus mobile advertising.
Aligning surface areas
This leads us to the first challenge, comparing different media in terms of their sizes. Route works on the basis of m² whereas typically screen sizes are measured across the diagonal in inches. As such, we need to convert Route sizes to inches squared in order to offer fair and true comparisons.
To convert from meters to inches a multiplier of 1,550 is applied to Route surface areas i.e. 1m²= 1,550 inches².
Setting a base line
Our multiplier uses the surface area of ‘Mobile’ as our start point. We therefore need to set the size of a typical mobile screen.
In this instance we have taken the median screen size from those listed on screensiz.es. Admittedly this is not an inexhaustible source, but it does include various leading brands of smartphones and seems to work in setting a line in the sand. Taking the median value again, we have a typical mobile screen diagonal size of 5 inches.
We then use the Aspect Ratio to help determine the actual screen surface area. More about this calculation can be via the Omni Calculator site seen here.
The Basic formulas looks as follows:
- diagonal² = height² + width² – from pythagorean theorem
- AR = width / height – it’s aspect ratio
- area = width * height – area of a rectangle
Applying these calculations, we see that the average surface area of a mobile phone will be 10.7 in². This is somewhere between an iphone 7 and an iphone 7 plus.
But what of partial screen ads?
The analyses here uses the dimensions of the full mobile screen as our baseline. This therefore assumes that all mobile ads are full-screen takeovers, which they are for video but not for “traditional” digital display ads that only take up only a small portion of the viewport which in itself is only a segment of the full surface area.
The IAB’s 2018 Advertising revenue report suggests that just 15% of mobile revenue comes from video and hence will be full-screen. Other media such as TV and OOH make use of the full canvas.
Next, we need to align the other media.
Size of an average TV
Using information from the BARB Establishment survey (Q1 2019), we can determine the ‘typical’ main set TV size.
The data shows the proportions of people in GB who have different TV sizes in ranges of 10-inch bands. Applying a mid-point factor to each, it is possible to calculate the total TV sizes, and then divide this by the population to arrive at an average TV screen size of 40 inches.
Again, this is measured on the diagonal. We have assumed for the purposes of calculation that all TVs are in 16/9 aspect ratio format as per mobiles and calculated the surface area from there. This places the typical screen size at 684 in².
Size of an average ‘Newsbrand’
Here we take the dimensions from the website Papersizes.org. As with mobile, we assume all Newsbrand ads are of full-page size.
The dimensions of Broadsheets and Tabloids are included as separate inputs in our calculation. To generate a single average surface area we take the ABC figure for each daily newspaper and multiply this by the relevant paper size to give a total surface area figure per day.
We have assumed that the Daily Telegraph is the only newspaper of ‘Broadsheet’ size (23.5″ x 29.5″) with the remainder falling under ‘Tabloid’ size (11.0″ x 16.9″).
We then divide the total newsbrand surface area by the number of copies sold to generate the average size.
This equates to 209.5 inches squared.
Size of average tablet and desktop/laptop
These are again derived from screensiz.es data and again we assign the 16/9 aspect ratio. For both tablet and desktop/laptop we use the median value to derive the surface area.
Summary
So, there you have it, should Mary Meeker want to better account for the whole media landscape in her time/spend analyses, she should start by including all elements of the media mix into her calculation – i.e. add OOH in the mix.
From there, she could consider Dominic Mills’ plea to include the impact of the advertising that is being delivered by factoring in the size of the ads.
To do this, a simple scale factor could be applied using the relative surface areas for the ads being served (admittedly, this doesn’t help radio, but they have their own online multiplier, anyway). Were she to do this, her data would see OOH offer a wow factor of at least 68 times that of a mobile phone and potentially offer a scale factor of up to 251,599.
Euan MacKay is general manager at Route Research