T.O.P. FAQ
Geoff Mills [MFA, BA, GradDipBusAdmin, DipTeach]

T.O.P. stands for The Out-of-home People-meter and is a THEORETICAL MODEL for the quantitative AND qualitative measurement of consumer exposure to ALL FORMS of outdoor and out-of-home advertising. At its heart, this model uses simple line-of-sight proximity detection RF transmitter/receiver systems as the tool to measure exposure to advertising and combines this with GPS (Global Positioning Satellite) readings to corroborate and verify those measurements.

What makes T.O.P. unique is that it is the first model for out-of-home audience measurement that ONLY uses 'the physically recorded existence of actual exposures to advertising' as its basis for all audience metrics (projections, etc). To date, all other known proposals for outdoor audience measurement have been based on statistical estimates combining census data, recorded traffic flows at key locations, government population movement models, 'trip logs' completed by survey participants and location data recorded by GPS units attached to vehicles. As accurate as these 'statistical estimate' methods may or may not be, they all suffer from one major defect...advertisers don't trust them.

In a current survey being conducted on advertisers, advertising agents and media buyers in Australia, more than * 85% indicated they'd prefer an electronic 'people-meter' style method of recording audience measurement with less than * 15% preferring the latest in 'statistical estimate' methods. [*Results as at May 06 2003.]

This lack of credibility in 'statistical estimate' methods for measuring out-of-home advertising performance, is probably due to several factors...

1. Advertisers know that at the core of all television and radio ratings, is the fact that a measured number of survey participants were actually exposed to the advertising. Advertisers are familiar with and understand the concept of physically measuring the behaviour of a small but strategically selected sample group and they trust this method. At the core of the 'statistical estimate' models being proposed for Outdoor Audience measurement however, there is nothing more than a well calculated guess. (My apologies to my hard working colleagues who are doing an excellent job at Arbitron and ROAM, but as anyone involved in statistics will admit...unless you physically 'measure' at least a small sample of consumer behaviour...it will always be just a guess...irrespective of how much statistical and support information you use to improve the accuracy of that guess.)

2. By definition, the better and more accurate the statistical estimate method becomes the more complex it becomes. This complexity however, creates its own problems, because when advertisers can't understand the process used to provide the measurement, they are unlikely to trust it.

3. Because most of the popular 'statistical estimate' methods being proposed are based at least to some extent on 'vehicular traffic flow', the results of these methods are of less relevance to the mobile, transit, indoor and small-format sectors of the out-of-home advertising industry. This has tended to fracture the industry into groups who advocate for a particular method and groups who oppose that same method on the grounds that it represents some perceived bias against them. For this reason, it is unlikely that ALL stakeholders in the industry are ever likely to advocate for (or want to become involved in the financing of) any ONE statistical estimate method. This has led to the rise of one or two methods being financed by one or two large players which in turn further undermines the credibility of each method in the minds of advertisers...and ultimately does little to convince advertisers that the out-of-home industry has its act together.

T.O.P. on the other hand, relies purely on 'physically recorded exposures to advertising by the survey participants' and is as accurate for large-format billboards as it is for street furniture, mobile, transit, indoor and small-format plant. T.O.P. uses virtually the same method and process as the already trusted television and radio people meter survey method (albeit with completely different technology).

T.O.P. is referred to as a 'theoretical model' because though it is believed that the methods and processes developed by Mills are technically and economically viable, so far T.O.P. exists only on paper. To convert T.O.P. from a theoretical model into a 'real life option' will require field trials and testing of hardware and software and in particular, a testing process commonly referred to as 'proof of method'.

Nowhere. At the moment T.O.P. is only a theoretical model. Its authors have been enlisting the support of several major players within the Australian out-of-home industry to put together a steering committee to oversee and finance the all-important field testing and 'proof-of-method' of the model.

The initial field testing and 'proof of method' could probably be achieved for as little as $50,000 to $60,000 but the full market implementation is likely to cost as much as a couple of million dollars per year across five or six cities in a market the size of Australia. Though this may sound like a formidable investment for the industry to support, it drops into perspective somewhat with the realisation that if because of the availability of a credible and accurate audience measurement service...the out-of-home advertising industry were to obtain even a one percent increase in marketshare...the result would be a 3,200% Return On Investment for the industry. [See How will T.O.P. be financed? later in this FAQ for details of these calculations.]

T.O.P. was developed by Geoff Mills, a consultant lecturer in Marketing and Technology at La Trobe University - Brisbane Campus.

Mills is a former General Manager of Commercial Production for the Nine Television Network in Brisbane, has senior management experience in both television and radio media as well as print and large-format media and currently heads his own marketing and management consultancy firm.

Mills is currently discussing the potential development of T.O.P. with several stakeholders in the Australian out-of-home advertising industry including...

• CODY
• ROAM
• buspak
• Australian Posters
• MMT
• OMD
• and several leading advertising agencies.

No, not really. There are GPS (Global Positioning Satellite) systems being trialled around the world, and most are discovering that 'vehicle based' GPS systems are reasonably accurate, but the small format 'personal GPS' devices are not accurate enough to provide consistently usable measurements of survey participants' exposure to advertising. [Susceptible to high levels of fade out - particularly indoors - and a typical operational accuracy of +/- 30 meters makes these 'small format' personal GPS devices suitable only for outdoor enthusiasts and low end consumer requirements.]

Vehicle-based GPS is much more accurate and stable, but of course only measures potential exposures to stationary 'street-accessible' media. This means that the same GPS system could not be used to include mobile, transit, indoor or small-format media which typically makes up between 20% to 30% of total out-of-home inventory.

To our knowledge, no one else has been testing simple line-of-sight proximity detection RF transmitter/receiver systems combined with GPS data to measure 'exposure' to ALL media...indoor and outdoor, stationary and moving, large and small!

The technology is broken up into four components...

The TOP Guns :: Small, low powered, low cost, waterproof miniature RF transmitter devices (just a little larger than the size of a matchbox) that transmit a 'beam' that closely approximates the 'viewable field' of the advertisement they are attached to. These devices continually transmit a stream of just two (2) bytes of encrypted data. The first byte contains the unique ID Key of the site/plant/media and the second byte contains a continuous timecode, defining the date and time. The component parts of these devices can be obtained quite inexpensively (from $10 to $20 each), but need to be acquired in LARGE numbers as every piece of out-of-home media theoretically needs to have its own TOP Gun configured, attached and powered. The good news here, is that with no internal moving parts, these devices should be virtually indestructible and have a shelf-life of up to six or seven years.

The TOP Watches :: These are the wearable 'recording devices' that are attached to every survey participant during the entire survey period. They consist of a miniature RF receiver, a SiRF low power GPS chipset with 16MB of Flash memory to store all data and a highly efficient battery capable of optimally providing power to the device for the full week of a typical survey duration. Though these devices will be considerably more expensive to acquire than the TOP Guns, only about one thousand or so will be required. (Because final decisions on hardware and software have not been made, it is impossible to estimate accurate procurement costs, but estimates range anywhere from as little $360 per unit to up to $690 per unit.)

The TOP Survey Analysis Software that correlates the provided data from all individual TOP Watches into the final 'reach, frequency, etc, results' for each survey.

The TOP Online Management Software, that allows participating stakeholders to extract and print out 'media data' as required from anywhere in the world.

How is 'qualitative data' measured by T.O.P.?

There are several options available to us to utilise the TOP Watch technology to ask survey respondents to indicate (for example) on a qualitative scale, how well they recall the recent 'media' they have been exposed to. The survey responses could range from zero (not recalled at all) to five (paid particular attention to).

Additional qualitative data could also be collected via the TOP Watch technology including, what action (if any) the survey participant intends to take in relation to any of the most recent media they have been exposed to.

Though we don't want to force or train survey participants to pay more attention to out-of-home media than they otherwise would, the value of impromptu 'qualitative surveys' like these would be priceless and give advertisers in out-of-home media a strategic advantage over advertisers in other media.

How are mobile and transit media audiences measured by T.O.P.?

The same way that stationary, indoor and outdoor media audiences are measured...by recording 'line-of-sight exposure hits' on TOP Watches worn by survey participants.

The TOP Guns that are used to transmit the unique ID Key and timecode are small enough and robust enough to be mounted on bus sides, bus backs, taxi backs, truck sides, etc.

These transmitter devices can be configured to transmit in a beam extending as little as 3 meters or up to a maximum of 150 meters in an arc of between 30° to 180° depending on the physical size of the advertisement.

Whenever a survey participant's TOP Watch comes in line-of-sight contact with such a beam, the 'hit' will be recorded (along with the duration of the exposure).

The technology requires 'line-of-sight' so if the survey participant is capable of 'seeing' the advertisement, the 'hit' will be recorded. If the beam is blocked by something solid, the hit will not be recorded. This means that TOP is a conservative measurement of legitimate exposures to advertising...and doesn't use 'guestimates'.

How accurate is the 'TOP Watch recording device' and will it suffer from dropouts if the user is a passenger in a vehicle and their hands are buried in their lap, etc?

Though the 'line of sight' characteristics of low-powered RF are an area for concern if the receiver is hidden behind the walls of a car or bus, etc, initial tests would suggest that as long as the signal is entering the vehicle, the 'proximity detection' will occur. (RF - like most forms of electromechanical energy - kind of 'bounces around' inside a metal box like a car, bus, train, etc.)

You can conduct this simple experiment yourself...using a remote control for a garage door. If the batteries of the remote control are at full power, the remote control will operate a door, even if it is activated from the lap of the driver. If the batteries are weak however, the signal is unlikely to reach the controller when activated from this position.

Understanding that a TOP Gun transmitter is somewhat more sophisticated than a garage door opener (and doesn't rely on batteries for its power supply)...the issue becomes more one of 'power management' to the transmitter, than of critical inadequacies with the technology.

Perhaps a more accurate analogy might be...
Imagine that you are driving down the road at night, there are no street lights and IF you were to turn your headlights off, you'd be plunged into instant darkness. Now also imagine that the interior of your vehicle is covered with one hundred small but highly sensitive light meters that beep the moment they detected ANY visible light. Now suddenly up ahead from the side of the road, someone shines a broadbeam spotlight directly at your oncoming car. The light from the spotlight would reflect around inside your vehicle and as you came closer and passed by, rays of light would bounce and spill throughout the cabin space. The result of such an experiment might be as you'd guess, that your car would be suddenly filled with the sound of dozens of little beeps as various light meters all around the interior of your car detected the passing light...in other words...MANY parts of the interior of the car, would receive at least some light. Now the purpose of this analogy, is to point out that if a highly sensitive light meter were placed on your lap as you drove past the spotlight, it is highly likely that it too would detect the light. Exactly the same physics can be applied to the TOP Gun and TOP Watch technology. So the issue of how well the TOP Watch will detect a transmission from a TOP Gun is more a function of the power of the signal being transmitted from the TOP Gun and the sensitivity of the TOP Watch receiving that signal, than of any critical inadequacies with the technology.

Suffice it to say, that one of the first 'control issues' that the researchers will focus on when conducting field tests and 'proof of method', is to determine the appropriate power level for TOP Gun transmitters and sensitivity level of TOP Watch receivers. Setting these levels too high, will result in 'hits' being recorded when the 'media' is not visible. Setting the levels too low, would result in problems like 'blind spots' inside vehicles, rain fade, etc.

Beats me? Probably because the traditional practice of recording 'trip logs' has been around for so long, a kind of myopia set in. Certainly the technology to allow this kind of 'electronic exposure measurement' has been around for many years...getting it all miniaturised into matchbox-sized transmitters and watch-sized receivers however, well these are more recent innovations.

How will T.O.P. be financed?

The financing of T.O.P. is something for the industry to collectively decide. It is assumed however, that as in the television and radio industry, an 'owner pays' approach would be the most logical. (Given that it's the owners who have the most to gain from the increased market share that a credible and advertiser-approved audience measurement service delivers.)

The collective costs of providing the service can be summarised as follows..

1. TOP Gun transmitter hardware, supply, configuration and installation ::
       Approximate cost :: Between $9 and $38 per unit
       Numbers required :: One unit per site/plant/media installation
2. TOP Watch receiver hardware, supply and configuration ::
       Approximate cost :: Between $360 and $690 per unit
       Note :: These costs could be amortised over a few years @ $145/an
       Numbers required :: 1,000 plus breakage/losses @ 5%/annum
3. TOP Management Services
       Approximate Cost :: Between $360,000 and $890,000 per annum
       (Depending on whether a local or national approach is adopted)
4. Survey Participant Payments ::
       Approximate Cost :: $52 per participant per survey week
       Numbers Required :: Between 8,000 and 26,000 per annum

Amortised over five years and based on an assumption that eventually 200,000 site/media/plant installations Australia wide would need to be equipped with TOP Guns, and that 26,000 consumers would be surveyed each year, the estimated costs to the industry per annum would be :: $2.68 million

RETURN ON INVESTMENT :: Putting this cost into perspective :: If the provision of qualified and advertiser-approved audience metrics were to gain the Australian out-of-home advertising industry a one percent ($86 million) increase in marketshare (of the annual Australian $8.6 billion advertising market) this would equate to a 3,200% ROI and a 33% increase in revenue for the sector (based on the $261 million revenue from 2002.)

In an open survey conducted on over 120 advertisers and agents throughout Australia since late 2002, respondents suggested that ad spending in  outdoor/out-of-home media would increase by 12.9% if credible and measurable audience metrics were provided. On current figures, this would equate to an increase in revenue for the sector of $33.5 million per annum (a healthy 1,000% ROI on an estimated maximum $3.3 million investment).

Given that media owners would probably like to see an estimated 'per-site' breakdown of costs, and given that the 'annual per-site' costs will be dependent on the total number of site installations (nationally) that are included in the survey process, (i.e. the greater the number of the sites, the lower the per-site cost because the total costs can be amortised over a greater number of sites), the following table provides an estimated annual cost per site at various site populations ::

Number of Site Installations included in the Survey	Total Amortised cost to provide service per annum	Amortised estimated cost per site per annum over 5 years
50,000	$2.2 million	$44.08
100,000	$2.36 million	$23.64
150,000	$2.52 million	$16.83
200,000	$2.68 million	$13.42
250,000	$2.84 million	$11.38
300,000	$3.0 million	$10.01
350,000	$3.16 million	$9.04
400,000	$3.33 million	$8.31

Example use of the above table :: Let's say there are 300,000 sites incorporated into the TOP Survey (i.e. there are 300,000 sites with TOP Guns), and a media owner had 1,000 sites included in the survey, the approximate cost to that media owner (using the simple formula above) would be 1,000 x $10.01 = $10,010 per annum.

Now of course it might be justifiably argued, that owners of large sites that are capable of earning thirty to fifty times the revenue of smaller sites, (a thousand supersites compared to a thousand taxi backs for example) should be carrying more of the financial burden...but the table above is offered only as an initial guide. It would be a relatively simple matter to produce a sliding scale based on the 'rate card' revenue of the various sites involved in the survey, so rather than breaking the 'per site' costs down as in the table above, the per-site costs could be broken down to be proportional to the earning capacity of each individual site.

Another issue to consider, is that some media owners will choose to equip ALL their sites, while other media owners (for sake of costs) might choose to only equip a small percentage of their sites and suggest to advertisers that all their other sites would be obtaining similar results. The former media owners might suggest that this is unfair and that media owners should submit to equipping at least some reasonable percentage of their sites before being able to benefit from the service.

A solution to these and other issues could be achieved with the establishment of a charge structure being made up of the following three components...

SUGGESTED CHARGE STRUCTURE ::

1. Each media owner would be responsible for the costs (e.g. $16.50 per unit) of acquiring their own pre-configured TOP Guns from the TOP Survey Administrators and installing their TOP Guns themselves  - which means that all TOP Guns will be consistently configured, the administrators will know EXACTLY how many TOP Gun sites each media owner possessed and media owners would only be paying for more complicated (and expensive) installations if they needed to do so for their own sites. 
   
   (Under such a system, media owners would apply to the TOP Survey Administrators for
        X number of Type A TOP Guns,
        Y number of Type B TOP Guns, etc. 
   where Type A, B, C, etc. are TOP Guns configured specifically for different forms of media...supersite, taxi-back, bus side, bus shelter, 24 Sheet, etc.) It is envisaged at this point, that the 'supply' costs for each TOP Gun would be the same, irrespective of what TYPE they were...the only difference between one type and another would be in the configuration of the depth and angle width of the of the transmission beam, though power supply issues between fixed media and mobile media MIGHT introduce a slight price variation if different component parts - transformers for example - are required.)
2. Each media owner would also be charged an Annual Fixed Fee (which would be the same for ALL media owners, e.g. this might be some nominal charge like $20,000 per annum, which would mean that even the smallest of media owners who included only a handful of sites in the survey would still have made a sizeable contribution to the operation costs of the service.)
3. Each media owner would finally be charged a sliding 'per-site charge' for each site (based on the 'rate card potential earning capacity' of each site and the total number of sites in the survey. - e.g. and I'm picking figures out of the air here...but sites that are only capable of earning up to $5,000 per annum in advertising revenue might be charged only $5 per site per annum, while sites that can earn up to $65,000 per annum might be charged at $65 per site per annum, etc.)

With such a Charging Structure, even the smallest of media owners would be making a reasonable contribution to the overall operating costs of the service, yet generally speaking, those who have the most to financially gain from the service will be contributing proportionally to its operating costs. The greater the number of sites included in the survey as a whole, the lower the 'per site' charges for ALL media owners.

Of course, final decisions about how the TOP Survey costs would be amortised and shared will be a matter for the industry as a group to decide...these thoughts are offered as unbiased suggestions only.

The developers of TOP will of course be contacting the OAAA in Australia to see if they'd be interested in participating in the administration of the project as an independent arbitrator, etc.

© 2003 Geoff Mills [MFA, BA, GradDipBusAdmin, DipTeach]