Posted by & filed under Transcripts.

The Search Insider Summit, 2014 held a discussion moderated by George Popstefanov – Founder and Chief Idea Officer, PMG, where the panelists discussed about the scope and prospects of Programmatic Buying. The members of the discsussion were:

You can view the video on Ustream


George : But I wanted to apologize. Unfortunately I did not get the IOWA or the British accent mikes, so you guys are stuck with a Greek microphone for the next 45 minutes. I do apologize. We’ve a great panel today. We are going to talk about the intersection between solution, programmatic. We’ve people from agency clients, providers as well, so before we get started, I would like to get the panel to introduce themselves, tell us a little bit about their Search background, experience with Programmatic and so on.

Jon: My name is Jon Kagan. I am the Senior Director, Search & Biddable Media from MARC, USA. Basically I’ve been around doing search way too long, 10 years next month, but part of my role now is oversee any type of media that could possibly be bid, so video, DSP, RTB, programmatic etc.

Frost: Only looked at their title; I didn’t know they had people in biddable display or biddable advertising their title. I am Frost Prioleau. I am CEO of where domains like platform, programmatic platform – basically that enables customers or advertisers to buy across display mobile video. In a way, it’s much like search as possible, so in the world of programmatic, we’ve been very focused on “How do we make programmatic as much like search and by enabling things like keyword level bidding, reporting, optimization?” and that sort of thing. So…and we see from the search side, I actually have never been in search per se, but we have customers who come to us from search. We also have lot of employees who come to us from the search side, so we see that transformation that both of them make or try to make.

Sid: Hi, I am Sid Shah and I head the Analytics at the Media Optimizer division of Adobe. We actually manage over $2.5 Billion spent across search, social and display. Actually I want to ask everyone a question: “Why did the search marketer increase the displacement?” – “Because it made a good impression!” [Laughter] It’s 9.30, what do you expect?

Frost: We have programmatic in a title and a joke in the same panel, is it not? [Laughter]

Cameron: My name is Cameron Urry. I work at Extra Space Storage. We are a national…like gods; we are the second largest provider of self-storage in North America. We have all over in 100 locations. I’ve been in search for 13 years, started with paid search and done a little bit everything in there. I thought I’ll come on the smartest panel here. I’m very excited to be here.

George: Awesome. So one of the favorite panel from yesterday was when we talked about keyword list targeting, and how we constantly played a cat and mouse game with Google and so for a search marketer like myself, it’s kind of exciting because programmatic kind of offers the thing to where you basically figure out who the user is, what intent they did, how much they’ve purchased, so it should be pretty exciting for a search marketer to make the move to programmatic. So why do we think search marketers would make good programmatic buyers?

Frost: First of all, search is not interesting or exciting! It is…

George: It is, for me!

Jon: For me, it’s being locked in a basement with no lights, sitting with spreadsheets for days…[Laughter] Now, one thing I’ve discovered is for search marketers, the skillset that is brought over for programmatic is eerily similar; the real-time bidding is something we are already accustomed to, optimizing and the automation in our programmatics, but it is not much more different than a bid role, with your basic platforms, so when we look at it, it’s just playing with different styled ads, is almost high level where I could say it to my…one …person I want to bring him in for.

Frost: Yeah, I mean, on our side, we see a major theme like brand marketers and search marketers really different and the big thing about search marketers is they are willing to let the data talk, right? So we’ve campaigns; I know we did like a acne product campaign and Justin Beiber ended being the search term and so that was a brand market, like “Why the hell are you searching on? Why the hell are you targeting people, you know, searching on or looking up Justin Beiber?” But the search guys could get it, like, “Okay, you know, this market…[Laughter]…” But the Search marketers will go, “Hey, that’s what the data is telling me, is that people whose search ended Justin Beiber are also converting on this acne product,” and you could guess why that might be the same audience doing that, and so they are much more willing to let…not have sort of preconceived notion, but let the data…you know, follow the data where the campaign will take them.

Sid: Yeah, I agree a lot with Frost. What I will just add is that for a search marketer to move into display, especially programmatic display, it really makes lot of sense because as a former search marketer, when I first saw display campaigns and talked to the, you know, display folks, I realized a lot of it is what I call ‘faith-based advertising’ – just hoping it works or otherwise on a spray and pray, so these metrics, that the discipline around metrics and the kind of thinking around the mathematics and the optimization kind of mindset, really translates well to a display and especially programmatic display.

Now, that said, certain very significant things that a search marketer has to do or rather shift in thinking to be successful in programmatic, so one of them for example is the importance of creative is huge, the fact that campaigns don’t last forever, is very important to recognize. Often a display campaign will be a couple of weeks; often times it is augmenting, you know, a TV ad. Search is bottom funnel usually, so you can put, make a nice creative, be done with it for six months, one year, don’t change it, nothing will happen. You try the same thing in display, your performance will tank! That is just one of the things to consider. There are a bunch of other things like what are the metrics you are going to look at? You can’t look at bottom funnel all the time for display, and then of course the whole attributions war starts, right, so what is the value of an impression, and those kind of questions come up.

If you’ve a very strong Search performance or intent mindset, you’ll definitely be quite distraught and disappointed for the first period you are trying to transition on, and not to mention, I know we’ve been dissing Google for not giving us enough information, but as a…I mean, I am an algorithms guy by training, so let me tell you, Google does a huge favor in search by cleaning out our log inventory, by simplifying it for us. Display is complicated. The amount of resources you need to actually do Ad Ops and just get the machine running is significant and then that’s when you begin to appreciate some of the things that Google brings to the table for search.

Cameron: Thanks Sid. So I am on a little bit of context. With Extra Space, what we’ve done is we were coming from like the search side, we were very data-driven, like pragmatic marketers, you could say. We were looking at, you know, lots of…we were looking at keywords, we were looking at the datas to help make the decisions. As we started digging more into our display, I mean, with our business, it was…it’s a very life-changing experience, when people need storage, so there is these life events that come in, you know, if it is a death or moving or something, right? So we’ve come at, like from at the search side is, “How can we get a storage, identifying where people are making those or where those life changes are happening? What’s happening on a..?” You know, there is just the search keywords; has lot of data that we can mind to get there, so as our search team just started looking at “How can we start bringing some more upper funnel?”, we started digging more into displays; this was like within the last year or so.

And our search team was the one that is sort of leading this, so we had, this is where our experience comes in, is we started seeing programmtics that way, that we can get the intent. I think, with the challenge that we started running into as search maketers, like you know, paid search people, in the data, we were looking at clicks, impressions, those are the…and really were managing like our campaigns based after just ones and zeros. As we started digging more into display, kind of what Sid was talking about is, it forces your search teams to start looking at your clicks and impressions as people and events, what is the message that’s going to make a difference to them when they see that Display so that you are not burning money, when you are doing your Display.

Then we see this really really awesome, like, where convergence, where it comes together with a awesome data and now the message is really important to what’s going, you know? So it’s been beneficial for our search teams to see that like ‘Now it’s more than just clicks,”, is people as a life events, that is bringing that message is from the top funnel, all the way down to what we are seeing at the search level, you know, has forced us to start understanding, like, “How do we carry that message from the beginning, all the way through when we are seeing again, you know, in the search, so that we are telling the whole story?” So I think it’s been really helpful.

George: So, since you’ve seen this transformation, what kind of skillset do you feel these search marketers have to acquire quickly to be able to be effective at running programmatic?

Cameron: Most…yeah, as far as search marketers, they come in with data, you know, they understand the statistics or the math behind it, and that’s awesome! The hard part is more like the soft skills; is the understanding, learning about the people. Every time we started doing more like, “Okay, we are going to do this campaign, for this segment, this audience group,” it was like, “Now, what do we say to them? What’s the experience we do over to them? What’s the landing page?” and then, create became a more important…We run that campaign, you know?

Some were successful, some were not, and it was like, “Okay, more and more about the customer. Okay, now let’s dig in with the more…we understand a little bit more about what’s going on, we’ve run a campaign, tested, how can we provide a better experience, like a true experience?” and then as always, going back to like, get to learn more about the customer, more and more about what’s going on, what’s impacting those buying decisions.

George: Awesome. Sid, from your perspective, what do you say?

Sid: Yeah, one thing to recognize is that search is a pull and display is a push. Think about it. So display comes earlier in the funnel, is more persuasive, the customer has not declared intent openly or very obviously. You’re trying to just test out intent. So the focus on display is actually in persuasion. In search, it is just to get the demand that’s already there, and find it in the most efficient way. There is some efficiency in display also; you can blast your ads, expensive ‘About the page’ ads and you know, waste your money; of course there is some efficiency, but really the focus of display is in that persuasiveness.

That’s why creative becomes important, that’s why alignment with your overall marketing campaign becomes important, especially the offline side. You know, TV and display have a connection more stronger than, you know…? TV and search do have a connection ofcourse, but the display side is earlier in the funnel, it’s about persuasion, it’s about engagement with your site or with the product, and search is about efficiently collecting the demand as soon as possible.

Frost: Yeah, we see…I mean, the biggest thing with being a search marketer, from the customer perspective, is getting them off click-only attribution, so that attribution thing, which I think you mentioned earlier Sid is, is lot of…they come over and you know, “Hey, we want to run a campaign for booking $800 flights and we are only going to give you, you know, click attribution or expensive TVs or only going to give you click attribution.” We just say, “Guys, we are not going to take that deal, because we are not going to win unless you are going to give some credit to display ad further up the funnel or something with full funnel attribution.” It’s really hard.

So that’s I mean…that’s when search marketers come over with that mindset, then I think, you know, you were saying, that you’re going to be disappointed, if you’re just going to apply the same click-only attribution metrics. So probably the biggest thing we see is they don’t open up their…broaden their mind about “Hey, yeah, these are the ads that actually can perhaps make a little bit of a difference in driving conversions,” and that’s it.

George: Oh, how about the ecosystem is pretty complex, right? You’ve the DMP, SSPs and so on, so how do you get them over the barrier to start thinking audience-based targeting and I think Google is doing that with retargeting atleast for search ads and now PLA retargeting is kind of trying to give us that ability to where search is more of based on audience-targeting rather than keyword-based targeting. So how do you get them through some of the barriers of understanding some of the things in the ecosystem like a DMP – tag management, audience management and so on?

Jon: I mean, when I look at my team and I look at what skillset they need to learn the most and it’s a concept of adapting to something you never worked with before. When you look at Search, it’s, “Alright, we do our targeting based off geography, keyword etc, and we’ll use the tools that we have available to do that,” but when we start moving into this programmatic methodology, we have demographic profile and we have interest, we have topic, we have all sorts of categories. It’s really getting the team to sit there and understand that we are no longer looking at for a search query; we are looking for ‘x’ person that fulfills this thing and the biggest skillset that I’m having to teach them is to understand that “Get away from what you know, and don’t look at your numbers. Look at the person out there that you’re trying to find.”

George: And then from a think perspective, we know search marketers could not make the full team, and so, what have you guys…have seen some of the team members. If you had one team member in the programmatic team that’s crucial to a success of a programmatic display or buying campaign, who will that be?

Jon: I very heavily rely on my analytics guy. He…I call him my profiler, if anyone wants criminal mind, do the criminal profile! His job is to say, “This profile, this person is the ideal concept; this is where you’re going to find them, feed this data back into our systems [BR marketing ?? 0:14:14]” and that’s..he is my bread and butter.

Frost: And so we take a little different approach here. I mean, so campaign setups are important, right, and so, you spend, you waste less money on campaigns setup. There is lots of things you can do to get, collect look-alike’s, you know, put tags down and say, “Hey, here’s a thousand people, let’s profile their search behavior and their slight hesitation they have before…,” so we start with the core, but we almost say, “Hey, let’s start running against pretty broad set and let the data talk, let the data…you know, optimize to the keywords, the times of days, the sites, all the third-party data that…” you know? So which creative is going to work well against users who searched on, which keyword 24 hours ago on this type of site, during the noon hour and just let the data drive that as opposed to going in with a lot of too many preconceived restrictions.

Sid: I’ll just begin by saying, “Do not underestimate the ad ops effort, supremely significant in display!” It’s…again, coming from search, if you break it down in 4 parts, right, so the technology, the algorithms, the ad ops and the analysts and the account management, right, the person managing the account, or the campaign manager, so I would say in search, most of it you can actually invest a lot in the technology and the algorithms and make a big leap. In display, it’s much less, so I would say between ad ops and the campaign management/analysis, you’re going to spend 60% of your time, 60%-70%.

The technology comes at about 30% and in search, it’s exactly the opposite, because once you have a campaign running between, you know, two or three search engines, you flip the switch on, and then you just focus on the automation and the optimization of the bids, but in display, that has a much lesser role.

Cameron: Yeah, I think for us, it was when we were getting into this initially, it was our data teams. It was, we have like around attribution, so it is matching up our data with the impression data we are…click-based attribution, that was a big one. As we saw it going through, right now, I would say like our statistician, like mathematics analytics and analytics people are very helpful in helping make sense of all this. As we’ve been moving forward into, you know, we are getting the learnings, we are understanding the math behind it, that’s now, you know, we kind of got that taken care. I would say right now what has been really helpful in doing is helping us reach further across outward channels.

We’ve an optimization team that are testing the website, they are owning the experience of the customer from all the way, like beginning to end – that team has been very helpful, you know, with us, with search people. You know, data people helping to learn the skills on how to reach the customer, because what we have to do is make a continuous experience. As soon as we reached that person, I mean, first, initially, it was a click-base. Now it’s starting with the impression – that experience, that person sees from the beginning, that first impression until they click and then until they convert, it should be telling a story from beginning to end. And so, our optimization team, there is a lot more brain-power with us and them to really craft a unique experience.

Frost: I would say…I mean, you can make it as complicated as you want, right? Here and there you can take the [inaudible 0:17:39] and say, “Hey, you got to understand everything before you move,” but there is, you know, lots of people, there are lots of companies in our space, who let people dip into the water quickly, run some campaigns and figure out how much they want to invest after that, so the transit company will come from search and say, “Hey, let’s start. We did hear about programmatic; we got to start doing it.” They start with some small budgets and then they could figure out, I think, everyone says, “Oh, we want to do this in-house.” Initially, you are going to be fully self-served and then we see, you know, as they get into it, more and more like, “Well, you do that, you do that…” and some of them do bring it all the way in-house and others decide with some…they just outsource it.

George: Well, with that, I mean, if let’s say our brand is not the programmatic and they want to take their first steps, what are some of the advice you guys would give them about where to start and please don’t say “Use that as a technology or tool!” That’s not an acceptable answer. [Laughter] So if they haven’t started, they’re looking to do either in-house or partnering, like what are some of the first steps you think they should do?

Jon: I want to…usually whenever we start a programmatic team, there is…the first concept is literally the word ‘team’. Being out of New York, we are able to quote people with already programmatic experience, but the reality is, not everyone can do that, so we always look to…we always find analytics personnel that have, goes across search, display and paid social because they always are going to be the first key and then what we have them do is we have them shadow the display team, who were in the [count ?? 0:19:00] of learning the targeting from there and then from that we pick one and hit, we have them teach the next person the same concepts, then you slowly build the teams because essentially we can’t be replaced by technology yet! So always first approach is build a team and knows how to do it, before you start doing, because otherwise the results are going to be pretty ugly.

Frost: Solid advice coming from search, especially just works from the bottom of the funnel, so I do think search is primarily the bottom funnel tactic. Next, the first piece in display to start is site retargeting, frankly. It’s been typically the highest ROI type of display and mobile programmatic, for the most part, so get pixels down, start retargeting the people who’ve been to your site, then as you go up the funnel, typically we say ‘socially targeting to next’ – that’s people who’s shown some intent and so that would be next, and then beyond that, you get into, as you really get into the point where you are happy with the way those are running, you’re trying to drive more people, funnel, try and drive more people to your site, so that you can site-retarget them, right? Then you look at more demographic targeting, third-party, all the more broader-based targeting.

Sid: I completely echo what Frost said. I think we’ve done a lot of retargeting for the ones who want to first try display or especially on the search mind to be expecting very strong performance. So this is a general law, right? The bigger the audience, the less qualified they are in generally, right, so when you’re going upper funnel, you can have huge reach with TV. They are probably, on average, have the least probability of conversion and as you go down the funnel, you have a very high probability of conversion with search, but the audience is smaller. So retargeting is right next to search, with slightly a bigger audience because these are people who haven’t converted; you know, you can apply a little bit more persuasion to display and get a conversion.

Just to add some data here, we ran some retargeting tests, right, so to see what’s the value for the retargeted impression. So we found it was actually pretty high. It’s about, on average between 20% and 25% of a conversion, is a retargeted impression. You’ll never get those numbers with regular display, but in retargeting, you do get that. The other thing I would say is as you scale your program, you know, don’t go any assumptions. I’ve myself had learned a lot of things just because the data said so. I will just give you one example. So if I tell you that, you know, look at ads above the fold and below the fold, you might think, “Oh, I will go only for above the fold. Who needs below the fold? It’s a waste of money.”

But what we are finding actually is that the premium you pay for above the fold is so much more than you know…or rather the below the fold price you pay is so much lower that even though the conversion rates are lower, it’s far more effective in terms of ROI. This is one example and there are so many dimensions to look at, time of day, day of week, you know, the number of times sites…you’ve visited the site and so on and so for every dimension, it can get very complicated, but don’t go with a lot of assumptions; try to just see what the data says and make decisions from thereon.

Cameron: Yeah, I think this is fantastic! I mean, retargeting will be a great place to start. For us, it was testing some assumptions. We feel like we’ve customers that, you know, this is the type of behavior we wanted to know from the data and it’s like, “Okay, I think this is a great audience. There is enough people to go in there; now I will start running some ads retargeting, see if we can bring it in.”

George: And I will say some…I think from my experience, how about tag management for…and then the audience management. I think we’ve a lot of clients that want to get advanced into programmatic and do that, but they do not have a good tag management in place or a good audience management in place and they are so quick to jump into third-party data and bringing that inward, even take advantage of the first-party data, so the advice I would probably share would be make sure that your tag management is in place, make sure you understand where your consumer is and you’ve a good first-party and manage them a solution, in-house.

So with that, especially for a [inaudible 0:22:53] that they have a lot of visibility, what are some of best practices you could share? I mean, you shared some of onward running, display below the fold, which, for a lot of luxury retail brands, they will kind of cringe at the thought of that, you know? They want…

Sid: I’ll just add little more in. I admit, at first, even I was like, “This doesn’t make any sense.” Like, I talked to a data scientist, talked to some very smart people in our algorithm team, and what they said was contrary in view. They said, “Well, think about it. If you are on a page and you are actually engaged with the page and you scroll down, you know, you are likely more engaged than someone just looking at the top of the page,” and Google came up with the viewability report, I think couple of weeks back and you know, the viewability of below the fold is not that bad. You know, by the definitions of the IAB, which is I think one second I think for the entire ad; I am not sure what’s the definition, it’s a pretty robust definition of viewability. And they found that below the fold is not that bad, in terms of as compared to above the fold, so yeah, just to add to that point.

George: What about from an algorithm perspective? I know with your efficient front-tier background…

Sid: So I was just going to say that. So coming from the algorithm side, yesterday when someone was saying, “Oh, you know, what if Google gave every cookie, every user to the advertiser?” and I was cringing. I said, “Do you have any idea what you are asking for? Do you the know the amount of infrastructure we’ve built on this Hadoop and High Systems you have had to build, just for display? Because yeah, fifty mille-seconds to decide whether you should want to bid on the ad or not and what bids should it be!” The amount of algorithmic complexity in display, it’s magnitude is higher than search – number one. Number two is, the inventory is all over the place. Most of the time, the analysts on my team spend, is actually figuring out which sites are making sense.

There is a lot of junk out there, in terms of inventory. So you know, as much as I have, you know, mixed things about things to say with Google, but in this case, I think they have done a…you know, as a favor, for search especially, which is like, you can be assured as good inventory. The other thing is Facebook. Facebook has great display inventory and the quality is pretty high. I found AdEx having high, good inventory, Rubicorn has good inventory, for example, so you should try different exchanges and see what works for you in terms of the quality of conversion and the quality of, you know, referring sites and so on in the exchange. Those are some of the things…

Frost: Yeah, actually, we tried not to optimize on exchanges; it’s just a blunt instrument, right? So AdEx has some great inventory; they’ve got some bad inventory. Apmex has great inventory, they have bad inventory, so we try and…

George: Mostly bad!

Frost: You said it! Okay, so the ability to optimize really more on, you know, domain or page level as opposed to just saying, “Hey, you know, one exchange is not…”

Sid: Well, we think we have good algorithms, but…[Laughter]: I said we have good algos, but you know…

Frost: That’s what algorithms do, as I figure it out.

Sid: Well, our philosophy is we try everything and let the numbers decide for us, so we try all the exchanges, we don’t…but I am saying if you are getting started and you have limited resources and you want to try one or two exchanges, these are some of the ones I have found that works, right? But if you have the time and resource, then I would say try all the exchanges and see what works for you; your data will tell you that.

Frost: Right. I would say “See what domains work” because may be…because if you just throw…we still have customers who come and say, “Hey, I only want to serve on this exchange.” We are just like, “Okay,” but, you know, you are going to look at…there are a bunch of good sites, may be great inventory, may be a better ROI, then it …average there. So yeah, it’s really let the data talk, right? So try, you know, to extend, to start off, if you can use existing look-alike data, that starts the core bit, they start off fairly broadly, and it’s really beyond human optimizing now, because really you’re looking at combination of this domain with this impression…sorry, this creative screen-size, so many search on this keyword, may be at, you know, at this time, like the ‘go’ and just you can imagine the combinations that need to take place, because one domain may be great for someone who searched on one keyword or particular creative, but not so good for another.

Sid: One more little tidbit I found, surprised me, may be for you guys it’s not surprising, but I found that the mobile impression percentage is massive! I didn’t expect to see like 30% of impressions I see around mobile devices now, which is…

Frost: Are you including tablets?

Sid: Well, mobile is 21%, tab is about 10% – that’s what I found. And the conversion rates are pretty good, I mean, higher than desktop in some of the advertisers I manage.

Frost: And click-through rates are very good, but they are not all…

Sid: I am talking conversion rates…

Frost: Yes.

George: Jon, from your perspective, as an agency, I mean, you guys kind of optimize different clients with different goals, different setups, what is one of the best practices you could share with us?

Jon: If I told you, I have to kill you! It literally goes back…it’s…for years, the scene at search marketer mindset is: “Look at the front-end metrics; they determine the back-end metrics.” And now, all my big thing is as you look at what you want to achieve, look at ‘x’ ROI, look at all your data, this placement, this site, this network, this audience target, that is what you are going to optimize for, going with your ‘CNN’ numbers from there and see… basically it’s no longer applying with the front-end metrics. We go in…after we identify the audience and our target, start looking at the back-end metrics instead and then play from there. Two, for so long we have always been obsessed with what we must suggest to CPC to bring it down.

To hell with it! It might be better click-through rate that might change things or it might be just tweaking our target of our group or time bids; no longer just us looking at front-end metrics any more! That’s a very generic statement, but that’s pretty much [inaudible 0:28:30].

George: Awesome. Where do you guys see programmatic going in 2015?

Jon: I see it eventually helping to cut my overhead costs. As horrible as that sounds, we are seeing a fast track migration. Initially, when we first brought it out, it was like, “We are going to build a programmatic team,” then it was, “Alright, we can immigrate programmatic in display teams together” and now we are starting to go down the mindset of “No, not everything is going to be programmatic,” but it’s heading in a direction where it is becoming the dominant force of display, and when you look at the person…know how to do actual tactics or leads on the search side, so what I openly see is taking my programmatic and my search, probably my paid social and my display teams together, pulling it all to one and then having people do double-duty kind of scenarios.

Frost: Yeah, I mean, got…lot of ways with inventory and data types, but apparently the biggest themes going on the programmatic now is like the quality. There have been a ton of press-out…you mentioned, you know, Google will suppress on bought traffic, unviewable ads etc etc, and so advertisers are smart, they are demanding that all the platforms, like, “Hey, I don’t want to spend money on your boughts or unviewable ads or what have you…” so there is a big flight to co-audience, it’s pretty intriguing because all the money that was spent on that bought traffic and unviewable ads is now getting pulled away from their various deals, and directing it to good traffic, right?

So we are seeing some increasing CPMs because of that, but overall, it’s great for the industry, because people are caring and really unless…really spend a lot of time to make sure that they are minimizing the amount spent on that traffic. And I see, you know, another thing, just that we see is really fascinating in this group, especially is, sort of, programmatic following search, right, so programmatic…, you know, search has…you know, big players in search, who’ve brought that in-house, who have seen big players in programmatic, now bringing that in-house, instinct talk from Rick yesterday on certain local, in how, you know, big players in search are doing a lot of localized campaigns and we are seeing a lot of that too with programmatics, so…

Sid: It’s funny you said programmatic becoming search, because I see the other way – search becoming programmatic, in some sense! Because, if you look at search evolution, first was the bid and the keyword matchtype and then it became device. Then you got time of dating, day of week – sounds very much like display at this point. So I think search will converge towards display and display will converge more towards search. I think what search will get is a lot of more dimensions that an advertiser can look at and get more control on their campaigns, and display will get a lot more transparency and convergence type metrics and more cleaner inventory that search campaigns tend to have more of.

So I think they are going to converge in that sense. Now, the other things I noticed is that like you are going to see more and more, I would say, the…let me phrase it differently. If you want to be a DIY programmatic display team, then your team will be pretty large. I think, as was mentioned, I think the overhead cost will reduce because you’re going to find more…better tools and better systems that actually can manage this fragmented ecosystem, which, of course, will consolidate over the next coming years. I don’t see that happening in the next six months to one year, but I do expect that next year will be the further shift in that trend. The final thing I will mention is I think right now, depending who you are, programmatics are about 30% to 35% of all display, I think that will probably go up to 50% in the next year.

Cameron: I think the way I see, like the way we are looking at the programmatic piece is with, like our programmatic and our search teams, it’s making us see a bigger picture. We are seeing now like, I think what we are seeing is like our social teams, our content teams and then really what we are trying to do is we are identifying audiences, is forcing our search teams to be more audience-focused, is helping our SEO team, then go also, “Okay, this is the content that we’re doing to reach these audiences.” Our social teams, “This is what we do when we post; this is how we are going to go out and you know, spend our Paid…like social money,” and then on pretty much everything that we do, it then becomes, everything is pixel, with the whole display and retargeting.

It’s like that whole experience now becomes one and the big umbrella that covers this whole piece is our search team, so like the very first impression all the way down to, you know, when they finally move in and want our facilities, it’s…I mean, really our search team can hit them at any point.I think our fairly two really big pieces are search and there is email, that can hit these, you know, our customers at any point throughout that whole, that buying experience.

So that’s..I think for us that’s one thing that we’re really focusing on, is how do we leverage, you know, the great capabilities that display programmatic is giving all the other channels that we’re serving in our company and like other places that we’re trying to buy customers, to really bring them along, not just when they actually hit the website, but after they call our call centre, we can push that data back into what’s going on with our display, and like our…in our search campaigns and then also to really help them move in, because with our business, it’s a local business, we don’t make the money to actually go into the store and sign a lease, right? So there is this big part between when they make a reservation online or you know, with our call centre to when they actually transact with us, that is a big opportunity.

George: Yeah, it’s a good point. I mean, we’ve been talking about programmatic in display sense, but you know, we’ve seen somebody like turn, announce that you can buy TV from their platform right now. I think with the FBX, kind of the intersection between social and then programmatic displays coming together and I think…so how do we follow about that? I think Cameron’s point, you know, we see programmatic as a way forward of, not just buying…you know, search is becoming a programmatic , but not just buying display; also buying potentially social programmatically and then TV and so on, how do you guys see that in the next couple of years play out?

Frost: Yeah, I mean, definitely display has been first, but most of mobile inventory or last of the mobile inventory is programmatically traded over RTB, video. Video has been a little different. I think the mobile and display sort of bought together; video has spawned a group of video only platforms on that side. Probably over time, those will go together also, and then yeah, social is a different animal. You have SBX which is very programmatic, but then they all have their API, you know, the ads API and all that, so we see those more and more, even though it’s not real-time bidding, getting integrated with programmatic platforms.

Sid: I think what you are seeing is the unsiloing of marketing. We now no longer are going to be search marketers and display marketers; we are going to be marketers – integrated marketers. It’s been a pipedream for decades; now it’s going to happen! So what I am trying to say is like if you can buy, let us say, the TV is still not online, right, so you can’t really buy and sell TV inventory. It’s going to happen though – display is already there, so social is there, search is there. As the so-called offline stuff comes online, you have to think about not just your performance metrics for how the overall campaign will do, how will you create, you know, interest and then how will you create interest in the actual purchasing action, so you’ve to think of it in the holistic sense.

How do you pulse your campaigns? What’s the right media flighting? So you want to show initially TV ads to create interest and then drive it with display and search and so on. All this stuff is right now theoretical; a lot of it is done by media planners and large agencies, but more and more you are going to find that you’ll have the ability to actually execute on this at the click of a button. I think these are the first steps you’re seeing, you know? Social and TV are intrinsically tied; I mean, the connection is tremendous. If you show a TV ad and then you show a social ad, the efficacy of the social ad in conversions just goes up multiple…like 5x, 10x. You could see the data just speaks for itself.

So I am actually surprised more and more marketers are not taking advantage of that. So often times the TV team and the social team don’t really talk as much, but that’s going to have to [unsilo 0:36:39].

George: Yeah. To be fair actually, Twitter has a TV to social sync, where you can sync in your TV buying with your social campaigns and it works quite well, makes …so I agree with you 100%. Jonathan, from your perspective?

Jon: I was going to try and say something, but I after listening to Sid, he literally just said everything and therefore, it’s…we are no longer looking at media channels. We are looking at a single operation and then we go to a concept of whenever you are optimizing one channel, it’s the greater good.

George: That’s perfect. Well, we have about 5 minutes and I want to make sure we have time for questions, so does anybody have any questions?

Mack: Hi, I am [Mack Goodman 0:37:27] from Yahoo. You know, we have a unique perspective being able to see the buy patterns and buy behaviors across both search and programmatic display. We are seeing a lot of the conversions happening in our native product right now, which is great, but we are still seeing primarily…you know, the search marketer is buying on the YBN network, display marketers buying on our exchange, on the Mplus, and we are seeing both buying in the native product, so we are seeing, sort of, you know, both of the marketers mixing into that. So you talked about the unsiloing of marketing and I am very curious, from the planning side and the budget side and the agency landscape, what needs to happen in the next couple of years for those budgets to stop being siloed as well?

Jon: It’s…it goes deeper than just that. Everyone here is a digital marketer, recognizes there is a large level of politics between traditional and digital. It’s…there is no way it runs! That is a nice way of saying it. [Laughter] It’s often like a gang fight, and it’s the concept of the agencies, the big advertisers in saying, “Alright, we recognize this trend to push forward to a single, holistic media style,” though always the first step is you need to make your budget fluid. Constantly I am going to my TV team to beg for money to run something on TubeMobile or a YouTube or I am going to the print side and telling them their careers are over, so if I may have the money for paid social? It’s something like a…it’s always I am going and begging for this money, where if you look at it, at the media, if it was all one holistic budget, we could start optimizing it altogether.

George: I would say as TV becomes more programmatic, which I feel like it’s going to be, I think for a lot of brand marketers, that’s way above the head, even just the word I said “programmatic TV buying” – there are probably going to be a loss, and so I think they would come more and more and we are seeing that happening. They are coming more and more to us, to ask for strategy and advice, because the days of going out to dinner and pre-buy for 2 quarters and stick with a media placements are over. It’s more about real-time adjusting to the, you know, real-time things that are happening in the…you know, from a TV or from a pop culture perspective and I think we will see more and more aligned budgets and so on.

Sid: I have seen these fights first-hand and it is interesting. Most large corporations will have a brand person and the performance digital person and they usually don’t get along! Because…and to be fair, again, I come from a search background, performance marketing, you will wait a week or month to may be two months, may be six months to see your…you just say you are doing mortgages, your lead cycle is six months, you will wait six months to see the efficacy of your marketing, right? But brand marketing can often take years to see the difference. You might just cut off your brand and especially for a large corporation, you will not see any difference for a year or two. The question is: Well, they ever see more last 18 months? Are you going to wait 18 months to see the difference?

So often what happens is the decisions made at the top, at the allocations between brand and the performance marketing or digital channel and so on is the function of the bias of that person, what that person’s background is. If they come from a brand side, they will make branding decisions more; they will allocate, they tend to over-allocate on the brand side and then the person coming from the performance side will tend to over-allocate on the performance side and the brand will then take a hit. Let’s not…to make it more complicated, also think about pricing decisions, that has a huge effect and the economy which people don’t account for in these allocations.

We won’t go there right now, but to your point, what could be done and what could make this better, is if the platforms we work with…like for example, Yahoo, can provide us more visibility and transparency into how these channels are interacting. The more information we have, the better decisions that can be made, because the more the data, the less the opinions.

George: Okay, we will move that on a…the panel is I think running out of time. Thank you very much.


That was really impressive. George, I think that was the first panel or group-speakers of anybody that actually cut themselves off on time! And I didn’t have to play the bad guy! So thank you. And I also want to thank Grant for ruining my hopes and dreams because we’d actually gotten to a day and a half to this conference and nobody said ‘Big Data’and I was actually really really..

Leave a Reply

Your email address will not be published. Required fields are marked *