Stefan Beckmann, Country Manager DACH at Infectious Media gives his take on whether RTB is an issue for CMOs.

It’s likely that your advertising budget has changed shape in the last few years, with more going online due to a new way of buying. As a CMO you may not be excited by the underlying technology, but the efficiencies of real-time bidding (RTB) could answer one of the perpetual marketing challenges–how to do more with less.

Online display spending keeps growing and RTB is fuelling this growth. This is seen in media agency Zenith’s September figures, which forecast Internet advertising to grow by 15.1% in 2013, with display advertising the fastest-growing category at 20%. The IDC’s recent “Real -Time Bidding 2011-2016” white paper shows RTB’s part in this, with RTB increasing its share of total display advertising spending from 5% to 20%, between 2011 and 2016.

The momentum can also be seen in leading advertisers moving spend into RTB. Matthew Turner, BSkyB director of online sales and marketing, said at a UK IAB event in May, “Sky is putting 35% of its display budget through RTB, and there is no reason it won’t be 50% by 2013.” At another event Bob Arnold, associate director for global digital strategy for Kellogg Company, said its shift to programmatic and RTB buying improved ROI by a factor of five.

In our own quarterly report into the trends in Europe, we have seen the marketplace developing. In the UK after price increases last year, prices have been stable in 2012, with inventory growth from the opening of a number of private marketplaces; France has witnessed increased inventory liquidity with the opening of premium inventory source LaPlaceMedia, and the announcement of new ad exchanges from eBay, Facebook and Orange; and in Germany an increase in inventory liquidity has been predicted from the increasing demand.

This growth is based on RTB’s ability to show the right person the right message at the right time. At its simplest, RTB is the most effective way of placing display ads online. It offers marketers the opportunity to bid for ads one by one (or impression by impression) instead of by the thousand, allowing total control over when an ad appears, the creative message, and what price is paid. By using this technology, advertisers can truly have a one-to-one dialogue with consumers.

The initial step in serving an advert is the “impression level decisioning”. When a user clicks onto a new Web page that includes an available ad slot, the publisher sends out a call for bids. Advertisers then place bids based upon the value they see in that individual impression. If your bid is the highest, you win and your ad is placed. It is this ability to say yes or no to an individual impression that eliminates waste. All this is done in less time than it takes to blink an eye.

Impression-level valuation means prospects, customers and basket abandoners can all be valued individually. You may choose to pay a premium price for someone who has visited your Web site, searched for a product and then left without converting to a sale; or indeed for someone who mirrors an existing customer, whereas a cold prospect may be less interesting to you and therefore engender a lower valuation.

A key benefit in RTB is the ability to serve “real-time creatives.” This allows you to customise the ad displayed for every viewer. Using data captured through Web site cookies and data points (such as weather feeds and geographical location), creatives can be built in real-time, customized by impression with a message automatically altered to suit the viewer.

As the campaign rolls out, the technology is constantly learning and using that learning to improve the next set of decisioning, valuation and creative decisions. It is a virtuous cycle in which variables are constantly optimized and campaign efficiency improved.

However, for this all to work RTB relies on data; the better the data is able to describe your target, the more effective your advertising. No one would dispute that the data they hold in-house from CRM systems, their own Web sites etc is the most valuable data they own. So using your own data can translate to a sustainable competitive advantage. RTB-enabled advertising becomes a CMO issue when we look to use this first-party data, in conjunction with other data sources, to create unique target segments based on attributes unavailable to competitors. Increasingly advertisers are looking at how to integrate insights from the data they wholly control to improve targeting and generate unprecedented results. For advertisers with large volumes of data, the search is on for partners who can create bespoke data integrations that protect customer data while unlocking insights to power advertising for all marketing objectives; be that customer acquisition; retention or up-sell.

Is RTB applicable to your advertising objectives? If you have an online advertising budget, your agency is probably already allocating some of it to RTB buying. For some advertisers, this is where it will stop. If your use of data is limited, the benefits of RTB will be based on cost efficiencies gained from optimizing spend against previous campaign activities, and workflow benefits for the agency around creative delivery and budget management. In this case your team is probably doing all you need it to, monitoring results and allocating spend to the partner best able to deliver efficiencies. But if your business holds data at any volume, be it on your customers, members or Web site visitors, the value RTB can deliver to your business is significantly increased, and the question becomes who can help you leverage your data to drive wider business success. A partner needs to be technically adept at integrating multiple data sources and manipulating “big data” to generate insight; have people with the right skill set, able to act on that insight for advertising campaigns that deliver results against business-led marketing objectives; and give your team transparency and control.

So is RTB a CMO issue? Perhaps the question is more “are you a data-driven CMO?” and if so, shouldn’t your advertising be data driven too?

Sylvain Deffay, Country Manager, France at Infectious Media gives his take on the impact of viewability on the French Display Market

As 2013 kicked off with the e-marketing conference in Paris, it seems that the importance of measuring online advertising viewability has impacted the French programmatic market. I was presenting along with several of my peers on the possibilities of viewability and where the technology is going, whilst Alenty presented jointly with AppNexus about their latest viewability app.

To date, viewability has been associated more with branding campaigns than performance, for obvious reasons. However, by ignoring viewability measurement in performance marketing, we are implying that the click remains the best measurement, and not the impression. It is time for us in France, with such a strong performance market, to explain and promote the efficiency of seen impressions in generating conversions, even without a click.

A measure of viewability can help us do this, and could not be more timely with the latest reports showing that, on average, anywhere from 30-50% of impressions are not viewed in standard run-of-network campaigns. The good news is we are now in a position to filter the real from the fake post impression conversions. Firstly, there is no need to account for post impression on unseen banners. Through comparing the uplift in conversions based on accumulated view-time, instead of just the usual frequency metric, each advertiser, and its trading desk, can now define which of the tracked post-impression conversions can be really considered as genuine conversions. This can be a great interim strategy to eliminate accounting for unseen impressions.

As viewability drives smarter measurement of performance, it entails a smarter buy from the trading desk. Many of us in France can now effectively measure the average viewability of adverts for campaigns, but only after the fact (a posteriori). Doing this takes a lot of manual work, however. By accumulating experience, site by site, publisher by publisher, the number of non-viewable impressions can be minimised when carefully selecting inventory sources. This will give campaigns marginally higher viewability, but has the related effect of excluding whole domains from a campaign, drastically narrowing available inventory and limiting targeting options.

The best value of measurement tools will come when you can automatically optimise campaigns to the predicted viewability of every single impression in real time. True automation will only come when this is integrated into the DSP, resulting in no effort from the trading desk to buy on viewability metrics.

The strength and scale of the performance offering in France has led many advertisers to view programmatic display buying through the lens of “last click wins”, with little room for the viewablity agenda. Viewability data offers the opportunity though, to not only to align with offline media measurement (characterised as eyeballs-on-ads), but also to provide more effective performance advertising campaigns, where all adverts can have a measured impact on conversion.

Whilst viewability data helps us all validate the effectiveness of our advertising campaigns, it also carries with it a groundswell that should change our industry in the long term.

Scroll down below to view our Infographic:

]]> http://infectiousdigital.com/blog/infographic-real-time-bidding-year-review-europe/feed/ http://infectiousdigital.com/blog/the-facebook-exchange-is-rtb-on-overdrive/ http://infectiousdigital.com/blog/the-facebook-exchange-is-rtb-on-overdrive/#comments Mon, 19 Nov 2012 17:02:46 +0000 Infectious Media http://infectiousdigital.com/blog/?p=775 Chris Stark, Product Director at Infectious Media, explains why Facebook’s recent launch of Facebook Exchange (FBX) is a massive opportunity which marketers should capitilise on.

For marketers, the Facebook advertising proposition is increasingly compelling. The site boasts 1B users globally which includes almost 65% of the UK online population. As a site built by users rather than publishers, Facebook is a rich source of consumer data. And because pages are designed individually for each user there is rarely a clash of ads, making Facebook a great brand-safe outlet.

Facebook launched its Ads Marketplace a few years ago to provide a channel that leverages its immense store of users profile data. Aimed at increasing activity within the Facebook environment, this display advertising gives marketers extremely precise targeting options. In combination with the personal connections that Facebook enables, marketers can raise awareness and keep their fan base engaged.

However Facebook Marketplace lacks the ability to reach the bottom of the sales funnel. So by supplying the same inventory via an open exchange model, Facebook’s recent launch of Facebook Exchange (FBX) is a massive opportunity for marketers with retargeting strategies. FBX does not make its own data available, but crucially it allows buyers to utilise their 1st and 3rd party data sets and bid optimisation technology to participate in the inventory auction.

Facebook is different from other inventory sources because its inventory is ‘always on’, with users accessing it regularly throughout the day. In contrast, other RTB inventory is often weighted towards the evening. As a result, any activity requiring high frequency touch points will benefit from being run on FBX; for example, basic retargeting, creative and message testing, and ‘bench testing’ 1st party segments.

With Facebook adverts there are no viewablity issues, even though the ad-size (100×72) is smaller than IAB standard ads, they are always visible on the page. Incompatibility is not a problem because the site provides a high standard of consistency and appropriateness to the user. For the marketer, this means that a larger proportion of ad spend is focused on media costs, and not on site qualification or viewability metrics. The reduced ‘overhead’ means better ROI for simple retargeting.

So how does this work in real life? For e-retailers FBX allows for rapid re-acquisition. Many customers do not complete a purchase after visiting an e-retailer site; they may abandon their cart, get distracted, or stop their product research. The objective here is to get their attention, remind them of the purchase intent, and encourage them to complete the transaction. But e-retailers face cutting through an excess of online noise, in a limited amount of time, to reach these customers effectively.

The question is how long does the purchase interest last after the site visit? Two days? Five? The best retargeting campaign is one that can reconnect within hours. A good percentage of customers will be on Facebook for a few hours each day, so can be reached while they are still considering the purchase. So for e-retailers, marketing through Facebook Exchange allows for site visitors to be re-acquired faster than ever before.

For hotel chains, people become prospects the moment they begin to plan a trip. This planning is often motivated by reading about holidays or seeing friends posting pictures of travel abroad on Facebook. The benefit of FBX is you can apply external data sources to decide on the best time to serve travel ads. For example, we apply weather data to capitalise on a user’s increased desire to travel when it’s raining or cold.

There is also a real benefit in this industry, in running hundreds of small, sophisticated tests of pricing and travel packages in combination with the external data. Not only can you find the right customers, but the right messaging and the right opportunity as well. There are rarely simple combinations of creative and data that work all the time, the scale of FBX allows for the speedy optimisation of campaigns, minimising the cost of acquiring new customers.

Infectious Media has many years’ experience working with Facebook, serving sophisticated awareness-building adverts on the Facebook Marketplace via our own integrated tool, Impression Desk Social. We use this experience to maximise the benefit that can be gained through the Facebook Exchange, combining detailed knowledge of your customers, our robust optimisation and analytical techniques, and creative that can adapt to engage the audience in the most effective way.

Our integrated approach of FBX, Facebook Marketplace and non-Facebook RTB, offers a ‘whole that is greater than the sum of its parts’. We define and test your target audience with Facebook Marketplace advertising. We then drive them at scale to your site using larger ad formats on external RTB inventory. This audience is then retargeted when they go back to Facebook via FBX, well within the effectiveness window, and with a message optimised to their interests. As the system learns, it becomes more effective, optimising over time to produce better and better ROIs.

A common mistake is to think of Facebook advertising as similar to social media marketing. It is more accurate to think of Facebook advertising as a vast source of inventory and data, allowing the benefits of scale and consistency to be brought to bear to meet advertiser goals, making it one of the most effective advertising channels.

Harley Norrgren studied Statistics at University College London and after a two year tenure is currently heading up Infectious Media’s Analytics Team. Being with Infectious throughout their entire working relationship with RTB he’s had the opportunity to watch the space develop since the start of RTB’s European adoption.

If you’re spending large portions of your media budget on brand activity, you’ll want to ensure that you’re getting the best performance for your money, yet the effect of brand activity is typically hard to measure as it poses a few difficult questions to advertisers:

• What metrics can we use for measuring brand engagement?
• How can we get access to user level data from predominantly offline channels?
• How can we integrate cross-channel data to get a full picture of the effect of our branding spend?

It is our proposition that buying via RTB on Ad Exchanges makes brand advertising measurable and increases the accuracy of any direct response path to conversion analysis and attribution modelling. We currently create bespoke branding metrics (in conjunction with our data partners) and marginal attribution models, which help to give advertisers better insight into the effect of their brand spend and facilitate optimisation towards branding goals. These bespoke metrics are called ‘Brand Units’, and are a standardised unit of any combination of different brand measurements, such as exposure time or brand related search uplift, which can either be a proxy for conversion or reflect other incomplete path brand objectives.

Exchange generated impression level data feeds provide unique opportunities for data providers to integrate their data with unprecedented granularity. We are currently working with data partners to provide us with impression level exposure time metrics, enabling us to optimise towards advertising face time and to buy display in terms of view duration rather than per impression. The combination of engagement, search uplift and on-site activity density means that we can measure brand effect on a per user basis, rather than resorting to surveys, and optimise campaigns against these metrics in real time.

The unique user level insights generated from brand activity can be coupled with post brand engagement and ultimately conversion/post-conversion activities, allowing us to fit brand activity directly into the conversion path, as well as attribution models, providing advertisers with more insight about the path to conversion than was ever available before from a single channel.

One of the main considerations about running brand activity through exchanges is the perceived ‘remnant’ status of most inventory available. It’s important to remember that just because an impression is remnant, it doesn’t mean it’s of poor quality. Through the use of ad exposure times we can identify and optimise towards the best value Cost per Face Time inventory sources, giving advertisers more clarity and better value than direct buying could. Furthermore, the presence of private exchanges and the market wide increasing adoption of ad exchanges means that more and more ‘premium’ inventory is made available every day. Via exchange buying, adverts can be purchased on premium inventory on a per-impression basis and made measurable enough to get a stronger indication of the real value of each advert shown.

These insights represent a large step forward for a communication goal which is generally hampered by industry standard attribution models. We hope that the unprecedented feasibility of marginal attribution models and access to user level data means that more opportunities can arise for branding activity to be bought on exchanges and start to drive a shift away from historically retargeting heavy DR budgets.

For more information please visit http://www.infectiousmedia.com/index.php?page=our-products

As other forms of online advertising develop, the role and effectiveness of display online banners has been a source of much debate. Research (Dreze and Hussherr 2003; Ilfeld and Winer 2002; Internet Advertising Bureau 1997; Sherman and Deighton 2001) has shown that exposure to banner advertising leads to increased brand awareness, purchase intention, and site visits but the relationship between advertising exposure and actual purchase is still uncertain.

The main problem is “how do we measure the effectiveness of a banner?” The first and easiest way is through exposure based metrics: computing the number of impressions served gives an idea of the campaign’s reach. However, many advertisers prefer to allocate budget based on the sales target of the campaign rather than purely on its reach. Since they are not directly linked to sales volume, the number of impressions served has not been considered a good indicator for display advertising effectiveness. The next measurement evolution for banners, was then to look at click-through rates (number of clicks per impression served), but why is a click any better than reach? After all a click simply means a user has visited the advertiser’s webpage, but we know only a small number of these visitors actually go on to buy.   We then come to CPA (Cost Per Action) commonly considered the most reliable indicator for the efficiency of an online advertising campaign.  It is my opinion that CPA has its place but that it shouldn’t be considered as a measurement panacea for all banner advertising.

In few years there has been a proliferation of online advertising vehicles, so that today we can choose from:
* Banner Ads
* Mobile
* Paid-search (PPC) and SEO
* E-mail
* Video
* Social Media

It is my contention that, each has a different function. The time gap between becoming aware of a product and buying it can be quite long, and along this path to purchase different display ad formats can perform better than others: Social media ads can spread awareness among friend networks: Video ads are more suitable in convincing someone of the attractiveness of a product; SEO optimization can enhance the likelihood of a direct conversion, but what about banners?

Have you ever seen this picture?

It was shot in Burkina-Faso, one of the poorest countries in Africa, and one of the poorest countries on the planet. There are several examples all over the World of similar Coca Cola boards placed in villages where people are so poor they cannot afford to buy water and food, not to mention Coke…so why does Coca Cola spend money on ads which have a likely conversion to purchase of almost 0%?  I believe the answer is that they are not aiming for conversions at this stage of the ad-process.

In order to understand and measure the effectiveness of an advertising campaign a number of advertising hierarchy effect models have been developed. The most basic and interesting in my opinion is the Lavidge and Steiner’s one (1961) since it links the different steps a customer takes with three main psychological stages (see diagram below).  The idea is that rather than directly jumping to a purchase, consumers have to fulfil each step before moving to the next psychological stage. The gap between one step and the next can be short but reaching the next psychological stage generally requires a longer time.

This is why certain ad types perform better than others during a specific stage or step. There isn’t an absolute “best ad”, but: banners tend to be better than SEO in the cognitive stage, while video ads perform better in the affective stage. Using CPA as the only indicator for the efficiency of an online campaign will actually maximize the likelihood of a purchase but only once that the user is already in the Behaviour Stage. Getting rid of banners is an easy way to lose a very big slice of the pie, represented by users that potentially could buy the product but without banners may not even know it exists. I believe the efficiency of banner ads can go even far beyond this model: in fact this is a model based on “consciousness”, since in every stage described something is happening at a conscious level…but what about our sub-conscious? In the last 10 years, studies on the brain activity has been widely quoted in the academic and marketing frameworks: As part of my academic studies I looked at Neuro-economics and the first thing you learn about the brain is that we do not know the brain We know what happens at a conscious level which is only 2% of our brain activity, but the remaining 98% is the “un-explored sub-conscious” which accounts for almost all of our decisions. When we choose something, we think we are evaluating data to take the best decision but actually most of our mind is already made-up and we are just fine-tuning what our sub-conscious has previously evaluated using a huge amounts of data (most of which are images and keywords) stored through perceptions in some “un-reachable” part of the brain. Data that we aren’t even aware of but that have had an impact upon us.

That’s why it is more likely that a man from Burkina-Faso will buy Coca Cola rather than any other soda. That’s why it is wrong and reductive to measure the efficiency of banner ads using a last-click attribution model (see the previous blog article from Harley on this issue) or to say expenditure on banner ads should be reduced.  For an advertiser wanting to maximize the efficiency of his/her campaigns it is important to understand the synergies between adverts. CPA with a last click-attribution model will often show SEO and PPC to be the best performing advertising vehicles. Using this logic, it then seems wise for advertisers to transfer all their funds to this kind of advertising. Nevertheless, by doing so, it is my conjecture that they will observe a drop in the overall conversions as they lose those customers in the Cognitive and Affective stages. SEO and PPC seem to perform better than other online advertising simply because current measurement practice is completely tailored to them. A smart advertiser will use mix of online advertising,  measured in different ways. CPA can be used at a “universal” indicator but then the attribution model must change appropriately to take into account how every different advert along the path to purchase has contributed to the actual sale: the focus has to be redirected to the big picture rather than on a single vehicle of advertisement.

to put it another way, strategic planning is pivotal to maximize the efficiency of a campaign: you can build the most beautiful table with a surface of ebony but if you want the table to stand properly, you should not forget to take care of the legs as well.

Images of Big Brother spring easily into the minds of a generation brought up on endless dystopia novels. We feel surrounded by governments whose desire to know all about their people is exceeded only by their fiendish organisation and ability to sift through reams of data almost instantaneously. Stories of leaked data abound*, growing ever more worrying as we realise just how much information we routinely put out into the world. And, much as we might like to say otherwise, this isn’t entirely unjustified: 90% of people have shared information with at least one site**. There is a lot of information out there about all of us. On the other hand, the sheer volume of data floating around is one of the very things that makes this sort of nightmare scenario so unlikely – the difficulty already involved in getting meaningful information about any given individual is only increased by the amount of noise that is now out there. Equally important is the fact that none of the information being made available is personally identifiable. It sounds like a small point, but the difference between the knowledge that Susie Johnstone was recently looking at flights to Italy and bought a bikini and the knowledge that computer 856076815463 did the same is huge.

Interestingly, the more people know about how the information about them is collected and what it is used for, the happier they are about it – after hearing details about behavioural advertising, 74% of consumers felt more comfortable with their data being used**. This kind of data and the ability to tailor the advertising served to someone’s needs and wants is what differentiates digital advertising from other forms, so it is vital that we reach some kind of consensus on what is and isn’t acceptable. The only way to do this is going to be opening a dialogue with consumers, asking their opinions and ascertaining exactly where their limits lie as well as making as much information as possible freely and easily available. Until consumers feel comfortable with the information we hold about them and how it is used, we will not be able to move forward and exploit the full potential of online advertising.

The recent EU ePrivacy directive heralds a change in the industry’s attitude to privacy. The requirement to obtain informed consent for all non-essential cookies will force advertisers into clear disclosure of the implications of a visit to their website. The difficulty lies in striking the appropriate balance – we do not want to adhere to the regulations at the expense of user experience. Over the next year, we will all need to work to reach a consensus on acceptable forms of consent, which best achieve this balance. As members of the IAB, Infectious Media is actively involved in policy development and best practice data usage in advertising, and we see this as a real opportunity for positive change.

*http://www.guardian.co.uk/technology/2011/apr/27/playstation-users-identity-theft-data-leak

** Statistics from IAB’s September 2009 study, in partnership with Olswang.

In a business revolved around display advertising, one of the essential parts of getting a campaign to perform is to build an appropriate Network (a Network is a list of sites on which we choose to display our campaign adverts on).
Remember the phrase “Location, Location, Location”! Well, location can, arguably, play a major role in determining the success of a business. E.g. A clothes shop located on a high street will attract more customers than if it were located miles away from civilisation.

The same concept applies for online advertising. To get the best out of our campaigns, we need to locate our adverts on sites with high reach to our target audience.

Now we understand the importance of a Network, let’s look at how to build one. How do we decide which sites to add to our Network?

What not to do….

1. Rely only on standardised industry classification methods
2. Judge a site by its name
3. Classify sites once
4. Label sites- black and white
5. Adopt a “one size fits all” policy where you use one Network for multiple campaigns

Building Networks may seem simple, however, if you don’t get the basics right, it may be harder than you think!

What to do…

Start by answering these questions

• What brand safety measures should you take?
• Who’s the target audience?
• What are you trying to achieve? (e.g Conversions, Brand Awareness…)

Always bear in mind the answers to these questions while building your network.
Search for sites you can bid on, use keywords and categories.

CHECK THE SITE! Checking the content of the site is the simplest and most effective way of deciding whether or not it is appropriate for the particular Network you are building. Remember, if you label a site inappropriate for one Network, it could still be appropriate for another Network.

Now you’ve got the basics down continue building the Network with some smarter insights. What sites worked well in the past when similar goals were required? Could they still be relevant? Why would the audience be likely to convert? What sites are they likely to visit because of this reason?

Remember, be brand safe, reach your audience, optimise towards your goals. Get these right and you’ll get a network that works!

Regardless of what you think about attribution modelling, we can all agree that whilst practically useful, last click models tend to see attribution as a yes or no decision. Is a model that negates all exposure¹  history, apart from the ultimate exposure, necessarily painting the most accurate and useful picture of which advertising spend is actually generating conversions?

Thanks to the increasing digitalisation of our industry, exposure level data is readily available for building much more sophisticated and informative models, which can inform more efficient budget allocation and increase ROIs for clients. Using statistical methods we can isolate the uplift on probability of conversion that each individual exposure has and therefore assign attribution on an ROI rather than binary basis: exchanging the notion of a single exposure driving a conversion with an each exposure has a small effect approach. Furthermore these models can be implemented at the start of a campaign and updated throughout the campaign at predetermined intervals, providing simultaneous analysis and reactive attribution which would reward the efforts of effective advertising rather than competing media buyers simply gaming some system for what is essentially a random conversion attribution.

Approaches like these do take more time and require more skill to implement than last click models but there are three compelling arguments for their use:

1. The potential return on investment and increased understanding of the conversion path for clients would outweigh the cost of implementation.
2. Advertisers are already implementing more complex attribution models for planning purposes.
3. The data is already available for use and the industry is coming to terms with dealing with data on a daily basis, change is coming so why should advertisers wait to be the last in line when they could be taking advantage of this data now?

To illustrate my point I’ve pulled some data from our platform on user behaviour under different retargeting conditions with a view to compare their behaviours against unexposed users.

We have chosen to examine the effect of aggressive retargeting versus less aggressive retargeting on a user’s propensity to make a return visit within 24 hours of the impression. Both types of retargeting will be compared against the natural return rate rather than against each other using a simple Chi2 test.

The aggressive retargeting (AR) group will be defined as users exposed to an advert within the same hour as the first site visit.

The less aggressive retargeting (LAR) group will be defined as users exposed to an advert between 24 and 48 hours after the first site visit.

For the AR group the results can be summarised as follows:

For the LAR group the results can be summarised as follows:

We found that there was no significant uplift on the return rate (p-value = 0.3354) for the AR group, but there was a significant uplift on the return rate (p-value = 1.327E-5) for the LAR group despite the overall return rate being lower. A last click model would have given attribution to the AR group and suggested that it performed better than the LAR group, but this is plainly not the case. An exposure level attribution model could discriminate between significant and insignificant exposures, assigning attribution to where behaviours were driven and painting a much more realistic picture of advertising effectiveness.

Furthermore, tying in campaign setup with attribution model insights generated on the client’s side has become easier than ever before: granular campaign targeting is available through most platforms so making the transition could be easily achieved on the buy side. On the client’s side, implementing and updating a multivariate attribution model based upon maybe billions of rows of data is no simple task. Yet the industry is already starting to rely on big data to inform their advertising decisions and when clients want the improved results others may be achieving they’ll be playing catch up if they don’t start experimenting now. So my advice is to start small and when the time comes for de-facto bespoke attribution modelling you’ll be ready for it.

Of course this is quite a small insight into a vast array of possible analyses that could inform attribution model specification and I’d be keen to hear your opinions on this.

¹ For more information on exposures and interactions, please see Measurement: The Elephant in the Attribution Room where we discuss measurement in attribution models.

Online was supposed to be easy. Actual, attributable intent and sales from advertising without having to spend the company’s pension fund on a piece of econometric analysis that you would neither understand nor trust. Last-click-wins gave us a benchmark to measure all our digital activity, allowing us to compare and contrast different channels and strategies in the same way.

But it was broken. Assuming that only the last click (or impression if no clicks were recorded) influenced a user to make a sale was not just wrong, it was misleading. Yet Display and Affiliate networks made huge sums building CPA businesses on this flawed methodology and when Search exploded on to the scene, no-one seemed to realise that Google had effectively stumbled upon the best exploitation of last-click-wins, using it to build the largest online advertising business in the world. Even now we have Criteo-style product retargeting and Affiliate voucher code sites that often snipe the last click like a seasoned eBay auction bidder, winning the best deals just before the timer runs out.

Times are changing, however. Marketers are now more savvy and are demanding a more accurate solution to the attribution problem. But before this can happen, we need to understand the complexities of online tracking a little more deeply.

When analysing online path to conversion data, you can typically find between 5 and 100 events that may or may not have influenced a customer before conversion. Since most adservers only record clicks and impressions, your conversion path will only include these metrics as events. Herein lies the first problem. Impressions are not ad views. In other words, just because your adserver has recorded an ad call, it doesn’t mean that the ad was actually served. Even if the ad was served, it maybe was not even seen, especially if served below the fold. In essence, although impressions appear more tangible, they are no more accountable than newspaper impacts.

The second problem is clicks. Clicks are seen as the only measure of engagement and intent online. But advertisers should be asking for more. Anyone who has done some click to page landing analysis will know you see anywhere from a 15% to 50% drop off. If your clickers aren’t even waiting until your landing page loads up, how engaged were they? Similarly, what about all the people that read the ad, maybe played with it a little, then returned to what they were doing before?

Clearly not all clicks/impressions are equal and on their own, do not provide us with enough information to form robust attribution models. Whilst you can create a model based on your current media activity, all it takes is some additional spend on some cheap, below the fold inventory or some incentivised click sites to skew attribution and devalue the sites that actually do generate true ROI. What is required is more information. Impressions need to be augmented with above / below the fold data, time and position on page and page context / quality. Clicks need to be supplanted by interactions and page landings. With these enhanced metrics we can begin to understand whether our adverts were indeed seen and how they engaged our target audience. Once we understand this, we can start to model it.

Luckily there are now many companies that can provide this data. The likes of AdXpose and Flashtalking can all provide interaction data and some of the impression tracking enhancements mentioned above. Page landings can already be recorded using existing pixel technology and many data companies such as Peer39 can provide page context. What stops us from running all these technologies across all campaigns now are the current incremental costs of such solutions as well as the technical difficulties in integrating this data with standard adserving data in one place.

Of course, having the capability to record this data is only half the story. Storing and analysing it at a time when most companies struggle to store and analyse their click and impression data is arguably a larger issue. Add to this the lack of statistical and analysis skills in most marketing departments, is it any wonder that marketers hide away from the problem and merely discuss the fact that last-click-wins needs to be improved but have no idea where to start?

Here is where RTB can help. RTB provides an environment that allows any company to exchange data with another, server to server, in order to better understand the impression being served. By providing the APIs to allow companies such as AdXpose, Peer39, etc. to integrate directly with DSPs and adexchanges, the integration problem goes away. Data can still be collected in two separate places but you have a common unique user id to match up the sets. Once integration is solved, costs can come down as more advertisers will take up the service introducing economies of scale.

This still leaves the data storage and analysis issue, but creating a fast and scalable storage and analysis infrastructure is not as difficult as it used to be. Companies such as Netezza and Greenplum can do it for you for a price. Alternatively, if you can afford the time to investigate and implement open-source platforms, solutions such as Hadoop and InfoBright can also work just as well.

2011 is going to be a year where these technologies combine to allow us to better understand all that our advertising is delivering. Soon there will be no excuse for marketers to stick with last-click-wins as we will be able to provide robust attribution models to support or oppose our hypotheses. When this happens, we will not only be able to better understand the value of events leading up to conversion, we will also open up the door to more branding activity being placed online.