Retailing and facial recognition, the future of DOOH?

Every single day it is obvious: there are more and more screens in stores, malls, store windows, gas stations, supermarkets... Is retailing the future of Digital Out Of Home (DOOH)? Screens are often installed where posters used to be, they steal the show. Switching from print to digital signage makes it possible to change messages on the fly, according to day part: it is a better business model for sales reps.
With screens, retailing can take advantage of facial recognition technology. The principle of this technology is to compare the features of people watching a screen with a large database of faces in order to determine their age and sex. Since no image is ever captured by the sensor (no picture or video taken), privacy is not an issue. To make things even more clear, the store informs customers and passers-by that there is a survey going on concerning contacts with the screen: they can easily opt-out...

Facial recognition is an unbiased technology, exactly what is needed for advertising and marketing research to qualify and measure audiences (cf. Allied Market Research). In certain situations, this technology allows a retailer to deliver an adapted advertisement, so called personalized, at the right time. But this should be used very cautiously: facial recognition could be perceived as creepy if used to recognize a shopper (as a former shopper or as a high-value shopper cf.  Consumers' attitude to retails in-store personalization tactics). Too much targeting kills marketing!

Why use screens in retailing?

Facial recognition by EIKEO (for demonstration only). 
From left to right:
Female, young adult ; Female, young adult ; Male, older adult
Number of contacts with the screen

Screens in retailing can be used both for branding and for point-of-sale promotion. Location of the screens in point-of-sale is of course decisive, it should be optimized according to the customer's path in the store.
As always with outdoor advertising, the golden rule in real-estate holds true: "location, location, location!". Advertising in retailing is based on context targeting: screens put advertising a few feet away from where products and services can be bought, right away.

Combining facial recognition with retailing brings a source of rich data for stores as well as for advertisers; also, it is a precious contribution to global in-store analytics. Facial recognition is not intrusive at all: no declaration, only observation.

What data is collected?

• How many contacts with the commercial in a given period of time
• Here we are talking about verified, real contacts; this is different from traditional OTS (Opportunity To See), used for lack of any better solution by outdoor advertising (OTS based on declarations of a sample of passers-by, intercepts).
• Qualified contacts
• Male / female, age brackets
• Context, proximity
• Which part of the store: the screen is precisely located, aisle, shelf, product (facing)
• Time of day and place
• Which city, which neighborhood (socio-economic data)
• When: day of the week, season, weather, sales, holidays, promotions, traffic
• Dwelling time 
• Exact time spent watching the screen (attention, attractivity)
• Deduplicated audience of a screen
• Cumulative audience and, if necessary, average frequency

Of course, one can correlate this data to other KPIs of the store (cf. SMS Store Traffic). Such brick-and-mortar data could become part of a DMP (first party data), deliver adapted ads to the consumers.
It is exactly what advertisers need and what they don't get currently from the usual measurements, whether TV, radio, print or outdoor, all of which are satisfied with opportunity to see, opportunity to hear, to read, etc. Of course, with facial recognition, there is no doubt about viewability, as opposed to web advertising.
Last but not least, facial recognition for retailing is perfect to check and compare creative solutions (A/B testing).

Interaction with a screen. Could beacons make it smarter?

Screens are all over the place. Indoor and outdoor. Out of home, they invade retail stores, subways, restaurants, malls, universities, museums, stadiums, train stations... For the time being, communication goes mostly one-way: from an organization, an advertiser, an administration to a passer-by, a prospect, a client. Advertisers use them for branding, to build and increase awareness. Of course, we would like these screens to be smart. Above all, we would like them to be able to gather data, smart data, and by so doing, become smarter. Data, only data, will make a screen smart. How?
How to make a screen, a shelf, street furniture, smart ? Adding Wi-fi, NFC, Bluetooth, QR...The screens used for Digital Out Of Home (DOOH), outdoor and indoors, are part of the Internet of Things. They can be monitored to save energy, manage content and advertising, measure audience (reach, demos, dwell time, frequency capping, etc.).

But, since communication with a screen takes place in a public situation, there are nevertheless some obvious limitations to interaction. Even if a message can be somehow personalized for a given passer-by according to gender, age (face recognition), or according to what she or he did just before or where (a precedent location), there remains a serious question of privacy and discretion.  There could also be a risk with crowds and security: no store, no transport facility wants too many people lining up in front of a screen.

Then come the Beacons for communication and location awareness. First Apple's iBeacon, and now Google's Eddystone. A beacon, as a wireless sensor, uses a battery to broadcast a signal around the beacon (via built-in antenna) ; that signal can be identified (Universal Unique IDentifier) and received by any device entering the range. The beacon leverages Bluetooth Low Energy to estimate the user proximity (distance between the responding app and the beacon). The beacons are therefore perfect for retail environments. The interaction is personal, intimate: the passer-by's device receives a notification and does or does not do what is suggested: looks at a product, purchases an article, shares information with a friend, etc.
You need an app designed to communicate with the beacon. You can configure apps to trigger events once you enter or leave the beacon's range (opt-in /opt-out). With beacons, advertisers can also broadcast the URL of their website (pushed in iOS Notification Center). In a retail store, when a customer picks up a product with a beacon or a sticker attached, the screen nearby shows a commercial about the product in question...

Beacons are bringing proximity and context to mobile, in a way that is simple and cheap. It works like geolocalization, allowing location-based actions. The beacon can trigger actions on the screen: suggest a coupon, show transit information, schedules, weather, traffic, nearby stores (maps), maps (path finding)... For instance, someone close to a screen, watching movie trailers may receive a coupon to visit the theater next door... Screens become smart.

Physical Web is on its way

N.B. Screens can also communicate with smartphones using Aware, a new Wi-fi standard that enables communication between Wi-Fi devices.

Out-of-Home TV: +9%

What is out-of-home TV (OOH)?
TV away-from-home includes TV-watching in bars, restaurants, hotels, stores, in other people's homes, in offices and work places, gyms, transportations, hospitals, waiting rooms... you name it.

Nielsen measured this TV consumption in Chicago (DMA N°3) with Portable People Meter technology (PPM). Nielsen acquired this measurement technology when buying Arbitron in 2013.
The three-month test, starting April 2014, combined data from an LPM (Local People Meter) panel and a PPM panel (adults 25-54, speaking English and Spanish).

The results are good news for TV networks and stations: the average increase coming from OOH TV is about 7 to 9% (a rating lift of 14% for sport programming); the increase is strong for daytime TV, although much less for prime time TV. But for the time being, OOH TV audience cannot be sold since it cannot be included in the current Nielsen metrics (NTI or NSI). It amounts to free audience for advertisers.

All the TV that fits? 
The question now is how to combine three kinds of TV consumption: in home, outside the home (OOH) and mobile (smartphones, tablets) which is both in and out of home?

NB : The PPM passive technology uses a device that tracks all kinds of TV exposure. It detects an inaudible signal (encoded by TV networks) but not actual viewing, which requires attention if not intention and could only be captured by a TV set with a sensor (as used for DOOH measurements. See Eikeo). The survey combines two definitions of TV audience: being in a room with a TV set on (People Meter) and hearing a TV signal (PPM). Figures for hearing are probably higher than for watching.

Programmatic ad buying for all video advertising markets

What is programmatic ad buying ? 
Instead of manual insertion orders, mails, RFPs, face-to-face human negotiations and lots of discussion, programmatic is mostly a world of data and automation, auctions, rules and market places. With traditional advertising, humans talk to humans (buyers to sales people) ; with programmatic advertising, machines talk to machines.
It all started with media buying for the Web (display ads). Programmatic takes over because it lowers transaction costs (information, bargaining, policing): less middle players are involved (but with new skills), meaning less friction, fewer mistakes, less waste. Only data-driven marketing can shorten the sales process, improve targeting efficiency and personalize ads (according to context, behavior, etc.). Programmatic ad buying will represent 42% of display ad buying in 2014 in the U.S.A. (source : Magna Global, September 2014).
What the Web started will soon affect TV planning and buying as well as Digital Out Of Home (DOOH). For the time being, the most current and well-known form of programmatic is Real Time Bidding (RTB), but there is also data-driven creative, real-time analytics (with GRPs), real time optimization, digital in-stream video ads, deals (Twixt, AppNexus)...

Programmatic for TV
In the USA, many TV networks are considering programmatic buying. ABC (national network owned by Disney) along with ABC Family will launch a trial this summer with FreeWheel, a company specialized in the monetization of video. FreeWheel was bought in March 2014 by Comcast. FreeWheel was originally started by ABC (through Steamboat Ventures, affiliate of Disney), DirecTV and TBS.
The trial will use FourFronts Programmatic, a software solution to connect SSP (ad sellers, supply) and DSP (ad buyers, demand). FreeWheel will provide inventory management and adserving. No linear inventory will be included, yet.
TV will slowly go programmatic, starting with cable or satellite channels, local stations or VOD before finally reaching the prime time upfront market (programmatic guaranteed). Cox Cable will use Clypd to sell commercials the programmatic way.
Note that in order to align with digital video, TV must first take viewability into account (selling only viewable impressions).
What needs to be tested first is the impact of automation on advertising value: even if the TV inventory is limited, TV networks fear a decline of advertising price as seen with Web programmatic.

Programmatic for Digital Out Of Home (DOOH) / Digital place-based media
To go programmatic, DOOH needs measurement techniques able to produce data to drive targeting. Data must include demos and geolocation (indoor positioning for subway or rail stations, malls or supermarkets) as well as "real world" context (aisles in store, etc.) to deliver relevant ads. In September, PumpTV (Australia) started automatically trading advertising (RTB) using a platform powered by Videology.
In-store analytics come with smartphones and sensors (iBeacon, Estimote, OnyxBeacon, nomi, IndoorAtlas, etc.). Note that Apple iOS 8 keeps Wi-Fi from identifying iPhones (MAC address randomization).
There is also face recognition: with Eikeo or Quividi, as well as with NEC FieldAnalyst or with myAudience, advertisers can describe and target audiences according to gender, age group.
Many companies are already implementing programmatic in DOOH:  Xaxis Places, Vistar Media, SN: Xchange (Kinetic), PlaceIQ, Vukunet, etc.
But, first, OOH needs accredited measurements (MRC, CESP) and reassuring privacy enforcement.

All video media are equal under programmatic
With data, all video media will embrace programmatic work flow. Digital video ad convergence means video ad campaigns (and GRPs) across all viewing platforms and devices.

Exemple : sociomantic (racheté par Tesco).