This ensures that the level of media investment is relevant to an advertiser's communication goals. The promise of programmatic advertising comes in five key points.

PROGRAMMATIC ADVERTISING - PROMISE 1:
TARGETING THE BEST AUDIENCE

THREE MAIN TYPES OF TARGETING :

• Behavioral targeting

Thanks to Internet users behavior and the identification of their interests, this type of targeting personalizes the content to which they are exposed.
It is based on 3 types of data :

• First party data all available “proprietary” data to an advertiser.
• Second party data : data from a partner which support and complement the proprietary data (first party data) already available for an advertiser.
• Third party data : data aggregated by specialized players of which they don't own.

This type of targeting is reffered as audience planning.

• Thematic targeting

This targeting type involves whitelists diffusion, according to the affinity of the selected sites and the targeted population.
For example : a sports equipment brand wants to target outdoor sports fans. So we could create whitelists of websites where this target is the most present through audience measurement tools such as Comscore. This type of targeting is also reffered as media planning.

• Contextual or semantic targeting

This involves displaying ad impressions on some digital environments depending on the page's keywords. Tools are used to scan the content of web pages and form keywords clusters (such as "winter sports", "mortgage" or "weekend destinations"). Contextual / semantic targeting can be particularly useful to prevent an advertiser from pusblishing his advertising messages on web pages by excluding certain key words. Some of web pages called "Brand Safe" could deal with sensitive topics (such as terrorism, an air accident or a health crisis).

PROGRAMMATIC ADVERTISING - PROMISE 2:
SEND THE RIGHT MESSAGE

Message's personalization according to the profile of the Internet user is one of the major challenges of the programmatic promise.
What are the different possibilities of message personalization in a programmatic advertising campaign ? What is the added value that can result from it depending on the devices ?

HERE ARE 6 POSSIBLE WAYS :

• Smart retargeting
  It personalizes an advertising creation from the browsing data of the Internet user. The most classic form consists in pushing back the product(s) viewed by the internet user or the one he puts in a basket on an e-commerce site + associated prices. The purchase has not been finalized, this encourages conversion.

• The web-to-store
  For an advertiser with physical stores it is possible to dynamically push the closest point of sale to the Internet user with a creation. The association of a dedicated promo code allow us to measure the effectiveness of the web-to-store campaign.

• Geolocation
  According to the geolocation of the Internet user, it personalizes the message they see which can increase the relevance of the speeches. An e-merchant will thus be able to highlight the best sellers by geographical area or a tour operator the top destinations by region to maximize the engagement of targeted Internet users.

• Yield management
  By the dynamic display of a product's price or service as a function of context data.

• Loyalty
  The use of CRM data make it possible to personalize the message for each customer of the database according to their history with a brand. For example : customer types : small, medium, large / last purchase recency / frequency of purchases / type of products or services etc.

• The pre-test of the creations
  It is recommended that AB systematically test multiple messages based on target audiences and delivery scenarios prior to wider delivery. This makes it possible to collect performance data related to each creative lead in real time and improve the ROI of campaigns.

DYNAMIC CREATIVE OPTIMIZATION (DCO):

To push the right message, you could use COD technologies which allow you to automatically personalize the advertising formats of an advertiser. With thousands of creative possibilities available, a unique set of creations can be formed! Some programmatic players have their own proprietary COD technology like Gamned !. However, companies with no media activity are now specialized in this service.

PROGRAMMATIC ADVERTISING - PROMISE 3 & 4:
BROADCAST AT THE RIGHT TIME AND ON THE RIGHT MEDIA

Broadcasting at the right time and on the right media requires an upstream understanding of the media uses made by the targeted Internet users. Which devices do they use in priority ? At what time of the day ?
The graphic below shows a typical example of changes in media consumption on the main digital media over the course of a day.

Depending on the time slot, it will be wise to accelerate the advertising pressure on mobile or on news sites.
The idea is to script the media plan as closely as possible to the target's consumption habits to gain relevance and increase engagement.

‍
Technologies synchronization also make it possible to sequence the broadcast in real time according to several events. For example : weather conditions, TV campaigns or even the geolocation of the Internet user. The use of a weather flow will allow an ice cream brand to communicate only when the temperature exceeds 25 degrees or for a winter sports brand to increase advertising pressure during snowfall.

PROGRAMMATIC ADVERTISING - PROMISE 5:
TO BUY AT THE RIGHT PRICE

A solid human expertise is the first guarantee of a media purchase made at the right price, that's why the media trader is so important in this variation in this programmatic promise.

Along with the media trader, most media buying players have their own algorithms. These algorithms automate some optimization actions with decision trees integrating several dozen or even hundreds of criteria. Each criterion is assigned with a YES / NO answer or a score that will guide the automatic decision and influence the purchase. You gain in efficiency by limiting tasks with low added value and increase your campaign performance.

THERE ARE 3 MAIN CATEGORIES OF CRITERIA :

• The cookie profiling criteria, linked to the cookie
  The decision to bid on a print or not answers a simple question. Does the cookie belong to the audience I want to target? If the answer is YES, we can then refine the targeting criterion according to the heat degree of the prospect vs. the core target.
  ‍
  Example: Has the cookie shown an intention to purchase the product in the last 5 days ? If the answer is YES, the algorithm decides to multiply the value of the bid by a certain coefficient to ensure printing. If the answer is NO, he can decide to apply the inverse coefficient to decrease the bid.

• The context profiling criteria, linked to the distribution context
  The decision-making process is the same, but the criteria relate to the advertising environment.
  ‍
  Example : Does the bid request come from a travel-themed premium site ? Depending on the answer, an auction multiplier coefficient applies.

• The placement criteria profiling, related to location and ad format
  They can intervene in the following way. Is my 300x250 format above the fold (the waterline) or below the fold ? The answer to this question will help guide the decision tree and then the auction.
  An algorithmic decision tree can therefore combine up to several hundred criteria. A few thousand decisions are made in a fraction of a second. The goal is to determine the right price to pay for a print media.
  However, it doesn't replace human intelligence, although this ensures the overall coherence of the media strategy. This also makes it possible to test innovative hypotheses which may prove to be significant performance gains.

BONUS - PROMISE 6:
TO MEASURE THE DEVICE'S IMPACT

In order to assess the relevance of the media device, it's essential to have appropriate measurement tools. These tools read the campaign KPIs corresponding to the advertiser's communication goals. The omnichannel understanding of the customer journey is one of the major challenges for brands in the era of digitalized touch points.

Solutions exist and coexist to measure its indicators (store visit, etc.).
In France, 85% of brands generate multi-channel revenue, which is revenue linked to an offline activity – at the physical point of sale – and online, via its desktop and/or mobile e-commerce site. 71% purchasing journeys are starting on mobile, however conversion rates are lower, 2/3 of French people pay attention to human contact when purchasing.
These figures can vary greatly depending on the type of goods or services consumed. A human contact will be almost mandatory for the finalization of a real estate loan while a train ticket can easily be bought on mobile.
Finally, a cross-channel customer is more likely to consume your brand's products or services vs. a single-channel customer.
Granular measurement of interactions with all these touch points is definitely essential as they make it possible to understand the whole customer journey and to fully appreciate the performance of its media mix.