Martin Kelly explains how forward-thinking brands are considering the weather (among other factors) when utilising an emerging wave of dynamic creative technologies for awareness activity
Real-time bidding (RTB) is a seismic shift in online display, and there is no doubt it and its associated technologies are changing the face of display media. With all the hype surrounding this area it is worthwhile highlighting what RTB facilitates in order to fully understand the opportunity it affords those working in the industry. In essence, RTB facilitates three key decisions:
1. Impression level decisioning
Previously a buyer had to buy in bulk (by the thousand) to reach the few ad impressions they actually wanted. Now, impressions can be purchased in multiples of one, vastly reducing wastage.
2. Impression level valuation
This gives a buyer the ability to place a different value on each and every ad impression. As a buyer there are many different types of targeting I can utilise, each of which I’m willing to pay a different amount for. RTB allows us to do this at a very granular, individual impression level.
3. Impression level creative
For each ad impression purchased we now have the opportunity to assemble and serve a bespoke creative. This is best illustrated by retargeting, which presents you with the last product you looked at on a retailer’s site, in a display ad when you visit a third party site. However, it is worth noting from the outset that bespoke creatives are not the sole domain of retargeting activity, as we will explore further on.
Taken together RTB technologies can realise the marketer’s dream of one-to-one marketing, by reaching the right person, at the right price and with the right message/offer. However in the majority of cases, current RTB-based ad campaigns are only taking into account two of these factors: decisioning and valuation. This delivers the marketer with the right person, at the right price, but with the wrong creative message. Unfortunately, with any of these three elements absent an opportunity has been missed and campaigns optimised in this manner often see very little uplift in performance.
In our experience the factor that delivers the biggest impact, but paradoxically is least explored, is the impression level creative. The full scope of real-time bidding for display advertising has in many ways been eclipsed by the success of cost-per-click retargeting companies as it offers unparalleled return on investment. It’s a simple but hugely effective solution, targeting the right user (previous site visitors) at the right price with the right creative (the last product they viewed on a retailer’s website). Retargeting is a clear example of how display can be transformed in to a new, much more dynamic and data-driven channel, though it is limited in its ability to reach targets at scale.
Moving up the funnel
Retargeting should be a given on any media plan but ultimately it is limited, as the advertiser is reaching customers who already have a high brand awareness and have demonstrated a purchase intent (having visited the site and viewed specific products). No advertiser can survive simply by retargeting, in the same way that no search campaign will only involve bidding on brand terms even though they yield a high return on investment.
The key going forward is to utilise these technologies further ‘up the funnel’, to customise messages at scale to audiences that have not yet visited your website. This is where the most forward-thinking brands are starting to focus their attention. From a technology perspective RTB allows us to dynamically generate a creative based on any data set that can be referenced when a bid request is made. This can be information stored on a cookie as well as non-cookie based information. This article focuses on non-cookie based information, as this is the key to unlocking the scale needed at the upper end of the sales and marketing funnel.
When a site sends out a request for bids on an available ad slot, it includes a number of freely available variables that can be used to customise the creative eventually delivered, such as the URL of the site the ad is appearing on, the exchange content classification and the IP address.
While this seems like information more suited to formulating bids, it can be used to dramatically increase performance by tailoring the creative. As an example, we have used IP address information to determine the broad geography of where the ad is being served on behalf of a-deal-of-the-day advertiser. The creative functions by pulling in the offer of the day that is local to where the ad is being served. Compared to non-dynamic adverts used as a control group we saw a 250 per cent increase in response rates for this client.
Take the weather with you
Whilst there are only a limited number of data points freely available on the bid request, there are many others available on the market. For example, semantic companies such as Peer 39 and Crystal Semantics provide thousands of granular content classifications accessed via an API, which can be used to tailor a real-time creative. Which data to use will depend on the need - if you are a software company, the knowledge that you are serving ads on business focussed SQL content v consumer, browser-related content could be used to tailor the creative content of your ad.
Infectious Media integrate data sets, then using advanced analytics, proprietary targeting and optimisation tools deliver insight which drives client's campaigns. One interesting general trend we’ve seen is a huge variance in performance based on weather conditions, not just for the UK as a whole but at a more granular level using IP addresses to determine region. We have developed the ability to use a regional weather feed into our platform and use this when making bid decisions, pricing decisions and also creative decisions, so our creative could reference the fact that it’s sunny, raining, snowing or windy dynamically based on the forecast conditions for the region in which the ad is served.
Increasingly advertisers are exposing their own site data via API’s which we can tap into to tailor creatives. This is already widespread in search for factors including stock availability in retail, but the same principles can apply to display. Ultimately, in display we believe an advertiser’s own site data will choose the audience to buy against and what creative to show.
Marketers are limited only by their imagination in terms of how to use data to create targets. There can be no doubt that real-time creatives taking in multiple data points present a challenge for the creative teams but with performance uplifts of up to 1000 per cent it is a challenge worth meeting, and one creative teams will increasing be asked to address.
The changes affecting display advertising are fundamental and powerful. Data is increasingly the driving factor behind biding decisions and, as we have seen, campaign performance can only be enhanced if its use is extended to real-time creatives. There is, however, no one size fits all approach to the use of this dynamic creative personalisation. The advertiser, the sector they operate in and their campaign goals will determine what different data sets should be used to power creative decisioning. There is still a need for a good solid strategy around what is going to drive performance for an advertiser as well as the smart technology to make it happen.