26 September 2012 in ExchangeWire EMEA 12 Comments

Wayne Blodwell, Client Digital Partner at Universal McCann, 'Programmatic Buying – It needs to be split in two'

Wayne is a Client Digital Partner at UM International, working across all digital channels within the EMEA region for a variety of clients within differing industry verticals.

It has been really interesting to see such a high volume of words penned for the ‘brand buying coming into the programmatic buying space’ argument of late. An area of such high-interest for marketers, buyers and vendors, and where everyone ‘in the know’ knows it is an area that needs addressing.

It’s important to understand that there are a significantly greater number of brand marketers and creative bods within the Media Industry than performance media heroes (my guess would be 10x as many). Many reading this article will think that is the legacy of the ‘Mad Men’ era of advertising, however these guys are just as applicable today as they were in the 50s, yet the ecosystem they sit within has significantly changed.

I feel sorry for brand marketers. The ecosystem is markedly different, and you only have to look as far as the agency groups with trading desks. Agency staff are plugging away with their trading desk offering without really understanding the brand’s history, ethos, audience, objectives and the vertical that they sit within — but the brand marketer has some over-enthusiastic trading desk operator utter the words, “We operate across 1,000s of websites and can target banners to audiences that look like your existing audience and we buy this in a real-time auction to ensure that this the media is bought efficiently.” Does the brand marketer really care? The brand marketer gets what programmatic buying is, they just don’t want to do it — who can blame them?

I want to focus on one very simple point for this article which I think will help bring programmatic buying to the fore for brand marketers: The programmatic side of the Display industry needs to be split in TWO. Programmatic Branding and Programmatic Performance.

Currently, both sides exist within the majority of media agencies (often planners are required to work on both), but they certainly are not separated in the trading desk arms of the holding groups, or within the technology vendors (for the most part). These are the people who can unlock the Programmatic Branding opportunity within the space.

The main reason I believe that the two should exist separately of each other is quite simple: they both need to be executed in completely different ways, to be measured in completely different ways and to fulfil completely different business objectives. To provide a real-life use case: those of you based in the UK, and are X Factor viewers, (may not be many of you!) would know that the 8.30pm and 8.45pm ITV ad breaks this weekend saw the launch of the new John Lewis campaign. If you didn’t see it, you may have seen it fill up your Twitter and Facebook newsfeeds, as it generated a lot of buzz.

This is what brand advertising is; it’s about users viewing emotive advertising to make them feel a certain way about a brand and then taking to social media or word of mouth to discuss their feelings.

John Lewis shift product online, but they generate significantly more revenue from their high street shops. They take to TV advertising to get into the living room of a household and generate awareness/favourability and intent to purchase the products within their high street shops. Can a digital campaign really generate footfall within a John Lewis shop? We can match the reach/frequency of TV easily online, but can the advertising format match a TV ad (its not ALL about GRPs online — there are considerably more formats than a TV spot)?

If we really want to bring the big brands of the world into the programmatic buying space (lots of the big brands don’t actually sell products online) then they need to feel like there is a really good reason for them to be there, not to be clumped alongside performance advertisers and bought/optimised on the same objectives.

In essence, the separation of branding and performance needs to be far more conclusive across all sides of the programmatic ecosystem:

- Advertisers need to be more conclusive with their campaign objectives.
- Agencies need to significantly improve their planning to achieve better/more relevant objectives.
- Publishers need to package their ‘premium’ inventory, data and content separately away from the performance advertisers.
- Trading desks need to adapt their strategies and personnel to understand and achieve different goals.
- Brand measurement solutions need to be improved.

I also wanted to pass comment within this article on the comScore white paper entitled ‘The Economics of Online Advertising’ as it was certainly interesting, but it bypassed anything to do with client objectives. There has always been a misalignment of the objectives of buyers and sellers within digital media; buyers are tasked with buying the inventory at the lowest price possible, particularly for performance campaigns (this is often how business is won and lost – inherited from TV-buying) and this obviously isn’t in the seller’s interest as they make less ROI on their ad space. What happens? More and more adverts are added to the page in low-viewability positions to offset this loss in ROI. However if performance clients can get this to work for their campaigns, then why change it? The way performance campaigns are executed is not going to change any time soon (it would have happened two years ago if it was going to).

How about publishers start to get remunerated through programmatic branding because they are selling all their ads on the same page to that one advertiser in highly-visible positions (user takeover), or the brand advertiser can purchase user-initiated expandable units on a CPM model, or the advertiser can purchase some exclusive relevant data, or even the holy grail – the advertiser can target an ad on a premium publishers website, mobile app and TV station. All of this creates exciting brand experiences for a user (I should caveat that the creative that is executed is equally important as the media buy for branding campaigns) which is ultimately what ‘programmatic branding’ should be about.

One of the major stumbling blocks we have with the theory of ‘programmatic branding’ is actually with the publishers themselves. Publishers have been on the wrong end of the race to the bottom for too long, but they haven’t helped themselves. Direct buys on premium pages can see placements bought for £10 CPM, yet these placements can be bought for £2 CPM in a real-time auction. It’s no wonder you hear buyers say, “Let’s not buy directly; let’s wait to see how much availability we can get through the exchanges.”

Publishers need to stop farming out their premium just so they sell out. Publishers need to create the scarcity of their premium ads and the demand will follow, particularly through ‘programmatic branding’. I would love to see a premium publisher openly say that any of their unsold ads would be shared out between their existing advertisers as added value. This is short term loss, but in my opinion would create long term gain.

Let’s split the programmatic space into two. It’s how we can move it forward.

Agency, Brand, Display, DSP, Publisher, RTB, SSP

Latest jobs on Exchangewire

  • Glen Calvert, Affectv CEO

    Good article Wayne, agree with a lot of the points, but think we’re a long way off from agencies being able to split the two out

  • http://twitter.com/Gu881n5 Paul Gubbins

    Great read, until our next twitter fuelled debate my friend ;-)

  • Alejandro Correa

    Good article! While I tend to agree that publishers are on the wrong end “of the race to the bottom,” and that creating scarcity will drive up prices, it seems that the real issue here (which you mention!) is measurement. Last view attribution, a cousin of last click attribution, which was brought to us by the same people who invented search, is not the right measuring stick for brand. If pubs start giving away their premium impressions as added value, advertisers may “find” that premium inventory “doesn’t work” unless they are smart enough to measure its effect down-funnel.

  • http://twitter.com/adlandpeddler humanrow

    We were the first company to 100% focus on video, and by definition branding, in the RTB space so I very much agree with your POV Wayne.

    We also advocate a move away from server based metrics when measuring branding campaigns, which is why we’ve just launched our real-time brand survey tool, baked directly into our media buying platform.

  • http://twitter.com/ProfessorHulk Professor Hulk

    Good article Wayne. Agree with some of your points others not so much.

    There is definitely two sides to the programmatic marketing opening up and over the last 6 months I’ve seen the gap widen. All good I believe and introducing brand spend into this space is good for everyone. Agree that in the programmatic markets current incarnation though that it may not be the best for the client or even the publisher.

    I don’t necessarily believe all the trading desks are the same though. We only need to look at two trading desks really and you can see stark differences in how they approach the market….vastly different.

    What I would say is yes keep the client in mind but perhaps speak to a few publishers on the matter. Ask them who is making the progress and who they can see making the biggest steps in the right direction towards enabling brand and how they do it.

    The points you raise about remuneration for publishers is a good one, putting money in their pockets with the brand spend currently within this space opens the doors to the opportunities you allude to for the future. The market has to prove itself first though for the publishers to open up…every publisher is afraid of losing direct spend to trading desks. For some it’s happened dependent which desks they chose to work with.

    I’ve seen £10+ cpms in RTB also for premium supply so I think it comes down to asking the question “Mr Publisher what does your bid distribution look like….who’s bidding 70% of all their bids at 25p and who’s bidding at £5+?”

    The ones bidding at 25p although saying they are operating brand and finding premium supply almost certainly not based of that bid, while perhaps the ones willing to pay the higher CPM’s truly are going after the audience and premium supply.

    Measurement is a valid point and cant help but feel that by saying “they need to improve” is a little weak – but I understand why you say it like that….because no one knows how to solve it! I wonder if anyone will do a piece on how they propose to introduce brand measurement to digital. I agree with you GRP’s are not the holy grail and viewability is a step forward but ultimately it just doesn’t seem to stack up to enough for me. I’ll hold my hands up and say “I know not the answer young padawan”.

    Perhaps if more brand money comes into this area we can see the brand programmatic market actually develop grow and answer the needs of the Mad Men style Brand Marketers. I think it’s possible.

    Good article Wayne. Appreciate you sharing your thoughts.

  • http://twitter.com/newsofdisplay News of Display

    About your comment on the “Economics of Online Advertising” whitepaper… for sure the buyer objective is to get prices down but as it is explained on page 15, if you manage as a publisher to have designs with highly ranks of validated impressions you have arguments to push prices up.

    Short term reaction on low prices is to create below the fold placements to get some extra money. Until now this has been working because large
    part of the performance for big publishers is on CPM but this might
    change. (it didnt happen until now as you say but at some point, with more programmatic buying, this might change as buyers can have the skill to decide the price of each impression).

    I think performance marketers do not have any magic tool to make bad
    placements work well.. Maybe some of the retargeters
    but remember that these also want priority and be above the fold.(They
    will always prefer a 1.5% CTR campaign than a 0.3% CTR campaign). That’s why they love ad exchanges, because they can actually pay a CPA price for a placement than on a direct relationship with the publisher they would ask for a performance-level CPM.

    Publishers should be aware of all this and maybe switch the whole strategy to a validated system with control on the RTB buyers.

  • Paul, ExchangeWire

    Playing devils advocate here, but if we measured correctly, ‘through the funnel’, then surely the need for silos would be made redundant?

  • http://www.facebook.com/wayneblodwell Wayne Blodwell

    I see where you’re going with this – it’s obviously really difficult to measure ‘through the funnel’ for the brands that are primarily selling product offline – typically post behaviour panels. I think brands metrics such as favourability are good proxies – and always have been. I mentioned in the post that brand marketers know about RTB, they understand it, but they don’t want to do it … and measurement is definitely an issue, but I think we need to actually recall on 100s of years of brand marketing and bring it into the programmatic space – this starts with user experiences + creative experiences – not just display banners.

    If we restrict digital advertising to the brands that only sell online, or the ones that are easiest to measure (performance brands), then I don’t think we can expect the display industry to grow very quickly? Feels like lots of the performance brands are capping out as it is?

  • http://www.facebook.com/wayneblodwell Wayne Blodwell

    I was meant to mention the comparison with TV – the beauty of TV is there is essentially one format (seconds vary obviously) for all marketers, and it’s what you do in that space that reaches out to the user if you’re a brand. How you do this online across hundreds of different formats and when the user is in a substantially different mind set is the difficult bit for brands. I think it can be done – but maybe I’m just optimistic :) .

  • Paul, ExchangeWire

    Daytime versus peak for TV? I’ve known TV buyers to buy peak for DRTV purposes?

  • http://www.facebook.com/wayneblodwell Wayne Blodwell

    Would be interesting to see if high impact placements online (not just 1 standard format display banner) would be a beneficial buy for a DR campaign. Would need to be measured correctly not just by mopping up on post-impression conversions through reach.

  • Petteri Vainikka, Enreach

    There is a lot of talk these days about ‘programmatic premium’ and ‘premium RTB’, yet this article is perhaps the first (note: by an agency brand marketer, not a RTB tech vendor) truly providing a solution proposal to the inherent multilateral conflict of interests embedded within the notion of premium publisher supply entering today’s lowest CPC obsessed RTB/exchange channel. Also, interestingly enough, the proposed splitting of the automated channel into two (one for CPC/DR retargeting and the other for branding) is already taking place… it simply has a long way to go.

    The CPC/DR retargeting version of the automated channel is what is today driving RTB demand volumes – and introducing new reach-like budgets to display (with many previously search-only marketers starting to also buy display directly or indirectly.) However, because of the business objectives of these buyers (inventory at lowest cost, standard formats, and largest scale), the success of this new market (or v2.0 of what we previously regarded as the ad network model) has made the exchanges a very unattractive place to sell premium publishers’ premium supply. (It is important to note here that also not all inventory from premium publishers is genuinely premium, i.e., should be regarded as branding inventory.) This market – and for the benefit of buy side – is where RTB truly delivers: the success criteria being lower CPC/CPA. This is also what the RTB protocol, along with exchange-based scalable inventory, was designed to do.

    The other programmatic channel has – paradoxically – very little to do with real-time or bidding, but simply leverages the streamlined trafficking etc. features of the automated channel, but in a very controlled environment. Most publishers – and vendors – refer to this as the Private Marketplace model. These two represent very different (placement and ad format) inventory, very different access rights (selected agencies only), a very different sales model (price floor driven, not bid driven – and with substantial KAM involvement), and are intended for very different buyers and buying objectives. This latter version of programmatic is for branding campaigns, with very little emphasis on real-time or bidding — but more on planning (right media environment, right ad formats, right audience), and brand response learning.

    In this emergent ‘programmatic branding’ channel, the focus of attention needs to be on how can additional advertiser – and publisher – value be created? (Lowering transaction costs with automation is not the same as creating new value.) Opposed to the CPC focus of the CPC/DR retargeting channel, here, the attention needs to be on creative analysis (which creative works for what audience segments?) and audience insights (which audience segments are clicking, who have the highest viewtime?). Solving these needs has little to do with RTB’s bidding dynamics, and pre-bid considerations, and much more with publisher side reporting abilities. Hence, the ball to grow display branding budgets – automated or direct – is actually on the publishers’ end of the court.