Sometimes you work towards a goal for a long time, indeed sometimes for a very long time. By which I don’t mean “Internet very long” (i.e. 2-3 weeks), but rather a period of months and years. You have the objective in front of you and you follow this path despite numerous obstacles, doubts, various bottlenecks, setbacks and warnings.

Today, nugg.ad has published a case study and at the same time announced the start of its Open Targeting Platform.

What’s it all about? The Open Targeting Platform is a combination of solutions, on which we have worked for a very long time.

1. A cooperative, open platform for publishers and agencies

The Open Targeting Platform provides different possibilities to control targeting campaigns across many publishers and adnetworks – a number of publishers can opt to coordinate a campaign themselves through their portfolio by offering overall frequency capping, for example – or an agency books a campaign involving numerous publishers and adnetworks.

What is special about the nugg.ad approach is the federal principle of the platform, i.e. it is always a loose cooperation between strong local partners (=publishers), by means of whose ad servers the campaigns are usually run (i.e. no central ad server is necessary). That is important, because it is the only way for publishers to retain control over their own inventory and bookings, which is often not the case in other forms of agency targeting. The agency also enjoys considerable advantages from this kind of targeting. First of all, it naturally does not need to build its own system or have its own tags installed (which would not make the publisher happy), etc. Above all, however, in this cooperative solution, in partnership with the agency, publishers and adnetworks can introduce their inventory and advertising space in a whole new manner. Simply by making a higher-quality inventory available, for example. Or by making certain placements possible, which would not be available by means of a network booking through the back door. Or by allowing high quality, interactive streaming advertisements, which must be integrated elaborately into their portfolio. Many restrictions of conventional agency targeting are avoided.

It is quite simple – such targeting campaigns just work better. And the L’Oréal case study published today with the Zed digital agency proves precisely that a high-quality branding campaign with a large range can be achieved with such an infrastructure, despite elaborate advertising materials!

“Considerably more range in the target group was achieved than we previously expected.” (Lothar Prison, Zed digital)

We consider this solution to be the best option to bring together all partners for an agency with their data and controlling opportunities. It means that agencies will be able to carry out impressive targeting campaigns (incidentally, also with their own target groups) – but above all, the publishers involved can demonstrate that they are capable of bringing their high-quality portfolios into the market in a combined and flexible manner, without the agencies having to resort to using Google or other across-the-board platforms (with all of the known disadvantages).

And of course we take on suggestions in the Online Behavioural Advertising Manifesto of the FOMA (Association of Media Agencies), which was prominent in demanding decentralised solutions for targeting with free access for all market participants – that is exactly the idea behind the Open Targeting Platform!

2. A platform for branding campaigns with completely new metrics and optimisation possibilities

We first introduced the idea at the dmmk 2007 to replace the much despised click as a metric with something new, particularly for branding campaigns. But why? There are methodical and economic reasons for doing so. From a methodical point of view, many – particularly with regard to branding campaigns – feel that the fact that someone clicks on advertising gives very little and at best an indirect indication about whether or not the campaign goal has been achieved. It so happens that campaigns are often directed at the “click in the head”. In other words, the user should remember a brand/product, think it is great and then, much later, choose it intuitively from the supermarket shelf. Clicks cannot show very clearly whether a campaign has achieved this goal – this rather requires conventional market research methods, which explicitly register the branding effect, such as brand awareness, brand affinity and purchase intention.

Marketers can use this information much better where the evaluation of the effect of their campaigns is concerned.

The task, therefore, was to make this market research effect measurement available in such a way that the technical click really could be replaced. This was only possible by means of a very consistent standardisation of the market research element. We worked together with our colleagues from d.core on this point, and the result was an extremely streamlined survey method, which could be automated, and which could accompany a campaign in an almost fully automatic manner.

The fact that the branding measurement now simply runs in parallel and is immediately available as a standardised index means that we can, firstly, also deliver standardised effect reports for the campaign (by which campaigns can also be compared with each other, etc.) and, secondly, it is even possible to optimise the campaign to its maximum branding effect, while it is running. To do so, the algorithm is simply fed and trained with the branding values.

In addition, with innovative contact class optimisation, which also makes it possible to avoid too low contact doses  (so-called frequency boosting), we have created the prerequisites to deliver the branding campaign in desired contact corridors, often of between 3 to 8 or 10 contacts, for example. This is extremely significant because contact doses under a threshold of 2-3 contacts are generally considered to be worthless, particularly in branding campaigns (read out press release for more insights on the matter).

We are convinced that a whole new generation of online campaigns can be carried out in the future with these two elements, with a focus on branding effects, while at the same time achieving large ranges and making optimum use of the high-quality environments of the participating publishers. The results of the first campaigns have proven us right and we are, of course, happy to have found, both in L’Oréal and Zed digital, such an extremely suitable case for trying out everything! And we are naturally very grateful to the participating advertisers for their trust in this cooperative approach and for their support in carrying out this campaign.

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

In the great interview with Eric Schmidt by Charlie Rose he was asked what kind of innovations we will see from Google in the future.

As one example Schmidt explained a very interesting service Google is working on – which can be used to explain the relationship between targeting and search.

The huge business Google built on his search business is in fact nothing else than a primitive but yet powerful and clear targeting approach. Show people ads according to their interests – which are derived from their searches.

The huge monetization Google sees from this is not only based on the fact that a lot of people are using the search engine and therefore Google has a lot of contacts and reach. It even more has to do with the fact that Google ads are much more relevant than other ads people see online.

So Targeting works – on a search page.

But only a minor part of peoples online behavior is on search machines – the majority of attention and online usage time is still (and will be) on content pages – because that’s what people are searching for.

So one of the most exciting questions is whether it will be possible to apply the same principles as Google did so successful on the search pages on the vast amount of content pages?

The answer is yes. And no.

No because one important link is missing: it’s not so easy to obtain peoples interest based on their online usage on content pages. It has to be derived from the usage, the content they read, the time they spent on which type of content and so on. This is less powerful than search because of two reasons. First content usage tells a drastically less direct story about the users interest. The content he or she is looking at could eventually be of very limited interest. Or – even worse from an algorithmic point of view – it’s relevance for this user could be difficult to obtain, for instance he could disagree with the content but still enjoy reading it. Or the reading could be related to a very short-term interest and tell almost nothing about this person in general. Or he could read about someone who died in a car that accidentially hit a horse without at all being interested in cars nor horses nor funerals at all…

So deriving interests and attitudes from peoples content usage is difficult. But of course it tells you a lot about someone if you’d be able to collect and analyze tons of webpages he visited over time.

And that’s the yes. Yes, it is possible to derive meaningful information about people online if you track what they are doing, especially what kind of content they are using more frequently than others.

But I wanted to use the Eric Schmidt answer above to explain why search and targeting are perhaps more related than people think.

Schmidt explains how a google search could deliver meaningful information in the future. He described how Google would us all available information about a person walking through New York on a holiday trip. It would use the location, time of day, speed of walking and other information that tells what this guy is actually doing and looking for. And then this future Google search would do implicit searches on his own and show up with useful information related to the building someone is just looking at for example. Because Google knows that this guy is interested in architecture.

"Imagine the situation where the person, the GPS, the phone, and this
constant searching creates a narrative stream.  It’s highly
personal and highly entertaining.  Entertain me."

So what does this tell us in terms of targeting? Well – what Schmidt described as the future of Google is actually what todays targeting systems are already doing when people browse the web.

With every single click on a content page the targeting system will do a “search” over the informations available for this user, combine it with the actual information of the current click and come back with a recommendation what would be useful for this guy.

Todays targeting systems already apply the implicit search Google is working on!

Of course this still is in an early stage and the relevancy of a search ad is still much higher and clearer to the user than any targeted display ad. But this is going to change – and then advertisers will see the same relevance and performance for their display ads as they are used to with their performance ads on Google Search today!

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

Targeting is applied science. You can fill bookshelves with knowledge about targeting, different approaches, technologies, strategies, statistics, key performance indicators, procedures, best practices, target groups, click through rates, conversion, media loss, case studies, etc., etc., etc.

But, as you already might have experienced: Top managers and decision makers are not interested in all these operative details at all. They just want to be “shown where the money is”! For top management you have to:

1.    Describe your company’s business in a simple model, using only the most important parameters
2.    Describe the qualitative impact of your solution on the parameters of the model
3.    Quantify the impact in relation to the required investment (Return-on-Investment calculation)

And this is how it looks like for Predictive Targeting:

1.    Model of your business

There are three major parameters for the monetization of a website/portfolio with advertising:
-    Available capacity (in Ad-Impressions per month)
-    Average realized fill rate (in % of capacity)
-    Average realized CPM (net!, in €)

It has proven to be practicable to split the whole inventory into three segments for a differentiated view:
-    Premium inventory (high CPM, e.g. homepages)
-    Working inventory (mid-range CPM, e.g. channels)
-    Long tail (low CPM, e.g. un-qualified content)

With these 3×3 values any website/portfolio can easily be described. Either qualitatively in a simple graph where the area below the curve equals the monetization:

Or quantitatively in form of an Excel table (fictive figures):

2.    Impact of Predictive Targeting

Predictive Targeting can have a lot of different positive impacts on your business. But there are two major ones under which most of the others can be subsumed:

I.    Increase of reach, e.g. in terms of …
-    increased fill rates
-    prolonged channels
-    a well monetised long tail
-    etc.

II.    Increase of CPM, not as targeting surcharges but e.g. in terms of …
-    new high-value products
-    increased wallet shares
-    increased eCPM
-    etc.

In the diagram the increase of reach leads to a shift of the graph to the right while increased CPMs shift the graph upwards:

The resulting arrow is what nugg.ad calls the Targeting Vector. It’s everything about predictive targeting packed into one single figure (a real KPI!). The area between the two curves finally is the increase of sales (“the money”) you get from Predictive Targeting!

3.    RoI Calculation

Now the impact needs to be quantified. Based on your individual business aims, benchmarks, experience and an analysis of weaknesses and strengths you make assumptions for the impact Predictive Targeting on each of the parameters (exemplary figures):

The result is a comprehensible benefit calculation at a reasonable effort.With these figures you are in the position to make a well-founded decision whether Predictive Targeting makes sense for your business or not. You can put the benefit in relation to the costs of a Predictive Targeting solution and make a Return-on-Investment calculation.Without such benefit calculation you couldn’t even say whether a solution is cheap or expensive. You would purely depend on gut instinct which is no good basis for budget discussions.

I would like to encourage you to take this kind of helicopter view at your business. I promise, it will be very illuminative for you!

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

Many people ask me what i think about the new interest based ads by Google that have been launched last week.

First of all it is important to note that the enourmous success and power that Google has comes from superior technology – not only in search but also for the monetization of search-result pages with according ads they seem to have the best technology and business model by far.

Having that in mind Googles move into the field of Behavioral Targeting could undoubtedly been perceived as a threat for our business as independent targeting suppliers. And honestly – of course it is a threat. On one hand… With its unbelievable distributed network of content sites and sites under measurement by other tools like Google analytics and Crome they know more than anybody else in the world about users browsing behavior. As long as the information for showing relevant ads can be derived from content-usage they seem to be in a very good position.

Googles Targeting can also be a threat for the whole industry as it comes with even more questions regarding user privacy. And a lot of questions are still open and need to be answered – Google should really start being transparent and cooperate with the European privacy guys (which we already did – it really is possible to convince them that targeting is not necessarily bad). As they now started to build their business around the data they collect and own much more agressively than ever before it is very very important to do that in line with data privacy legislation and even more – i really hope that they start playing that part of the game differently soon and try to build trust around what they do. Otherwise it is much more likely after this product launch that especially the european legislation will come around with really tight rules that could damage the industry as a whole.

OK – those are the threats. Where is the opportunity?

The opportunity clearly lies in the fact that Behavioral Targeting will grow much faster as everybody expected from now on. Targeting will be perceived as something serious and definitely relevant by Media Agencies, Advertisers and Publishers – which has not always been the case. Of course everyone is talking about Targeting since a long time. But in fact ad-campaigns based on advanced targeting methods are still not in the main line, especially in Europe. Targeting campaigns are still treated as special cases that need a high attention by trafficking and from the consulting. The typical targeting campaign is still a special case – something many agencies are trying to avoid if possible. Although it is not possible to get reliable figures about the percentage of campaigns that are run with targeting we think it is below 5% currently.

This will change now. Simply the fact that Google startet this service will change it.

There are two reasons why the targeting penetration will change dramatically in the next 18 months. One is quite obvious: Google will be succesful. Because they weren’t until now in their content network (ad-sense) which produced really bad results compared with the search-results product (ad-words). Ad-Sense will improve and so will Youtube ads (also low-performers now).

The second reason is not so obvious but perhaps even stronger for our industry: Targeting is now perceived as something serious. Something that might help Google to use it’s Doubleclick aquisition and enter the market of display ads. This will lead to more effort on the publishers side (and agencies as well) to invest in technologies and run their own targeting solutions. And that is the point where it definitely will help us.

Ok – but will this only be a short-term push? I am pretty sure that Google will not be able to own the display ad market with their technology – if the publishers play their game wise and smooth. I think the targeting game will end up like it did with Google news as well. There will be a strong and relevant service for long-tail targeting by Google. They will also enter some low end publishers of course – especially 2009 will help them to succeed. But they will never be able to offer high level targeting based on quality content. Too far away, not enough understanding, too much automization.

And there is a last point which even makes me feel comfortable if they would enter the display market more agressively than i expect. Google only offers what we call “simple” or “classical” behavioral targeting. Showing ads based on users browsing history. Of course the browsing history might be massive and contain a lot of details. But it will never show any interest in low fat products, frozen food or body care products. Because people do not surf for this kind of products. This is why we developed predictive targeting that derives additional information from surveys and other datasources. Perhaps some of the thousands of Google developers are working on something similar as well – i don’t know. But i would bet they are not. Because it’s not a question of technology but a combination of market research, machine learning and statistics. And this is not a field Google is in at all as far as i can see.

For me it looks like a bright future for independent targeting services – especially Predictive Targeting!

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

This is the brand new blog about Predictive Behavioral Targeting. It is run by the folks around nugg.ad but we will also invite guest authors and experts to contribute.
The blog will bring news and articles around targeting in online marketing with a special focus on predictive targeting technologies.
Of course we will also focus on how to make money and drive more succesful campaigns with predictive behavioral targeting!

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.