 It's the Cube, covering HPE Big Data Conference 2016. Now, here are your hosts, Dave Vellante and Paul Gillins. Welcome back to Boston everybody. Hashtag sees the data, a pretty unique hashtag that HPE has chosen for its Big Data Conferences. The fourth year of the HPE Big Data Conference, what used to be the Vertica User Conference, Big Data Conference. And the Cube has been here also four years in a row. So we've seen the evolution of the ecosystem. Evan Lin is here, he's a product manager. One of the companies in that ecosystem with visibility measures. Evan, welcome to the Cube, thanks for coming on. Great, thanks for having me. So tell us about your Boston-based company. Absolutely, so Visible Measures, we are a content advertising company and we specialize with working with brands and agencies and delivering their video ad content. We also do a lot of measurement services as well and that helps us pair over our media capabilities where we do offer a lot of data insights to our clients and helping them optimize their spend and their ad dollars. So we're really interested in this topic given that we do like eight zillion videos a week. So tell us more about sort of the company, what you guys sell, your value proposition, how you compete in the marketplace. Absolutely, so we do have our own hosted third-party digital video and display ad server as well as a viewability measurement service that operates across multiple different ad formats and video players via ad tag integration through publishers, web pages and video players. So essentially we do work with these brands and clients to deliver very custom online video ad campaigns based off whatever targeting they want, specific ad KPIs. And we pair all of the media that we run with them with a lot of data insights that we're collecting on the side in regards to what some of their competitors are doing, what their space is doing and what the trends in the marketplace are. So how do you target people? Let's say I want to target some particular persona. How does it work? So we actually work with a lot of companies like Nielsen and BluQai Escalate to grab their audience segments and then we sync that to, we basically use those data management platform, sorry, use those segments and target off of those. So when you say grab audience segments, you mean somehow you have a deal with Nielsen, you go into their database and say, okay, give us this many of your users and we can target them? Yep, that's exactly it. Okay, and you target them how? They're on the web somewhere and then you insert an ad? Yep, so we actually work with tens of thousands of different publishers and all these publishers have various metadata tags associated with the stair-table user that will visit those sites. So that's what we use for targeting. Now how about for individualization? Are there ways for you to know that it's me that's visiting that page using cookies or login information? How do you know that? Or is it just sort of a target persona? Definitely more towards our target persona. Element, that specific realm is slightly out of my domain. I handle sort of more of the data and analytics insights that we provide to our clients. Well, talk about those then. What kind of metrics do your customers care most about? Sure, absolutely. So visible measures actually started out as a measurement analytics company where we went out and we found branded video content across the web. A few years ago, those properties might have included Meta, Cafe, Fun Your Die, but today the more recognizable names are Facebook, YouTube, Instagram, of course. So what we go out is essentially in 2016 you've got a ton of video content spread across lots of big social platforms, smaller properties and it's really hard for a marketer to get an understanding of how their content is performing relative to itself and also relative to its competitors as well. So what we'll go out and we'll final this branded content for not only the clients that we work with, but for also tens of thousands of other brands that we don't work with and we'll go look at things like view counts, velocity of growth, social interactions occurring on these placements, comments, sentiment analysis and we'll sort of bundle all that together and present it to our customers. How about things like metrics on individual videos? You're collecting that as well. How long the video was watched? Yep, so on the stuff that we deliver on our own end, so the stuff that we actively deliver for our media, well we would absolutely collect events like quartile and decile beacon data. So on any given day we might be delivering a few million views for our clients and each one of these views has maybe 10 to 15 different unique events associated with them, whether it's click through, shares, tweets, as I mentioned, the decile or quartile beacon data. For the stuff that we're collecting more passively, so for the stuff that we're not running for our clients, we are still grabbing stats like view counts per duration, metadata tags, all that. So give us some free advice. I'm a marketer, I want to do a video campaign. What are some of the secrets of success? Sure, so part of it is what your goals are. Obviously there's a very simple solution of just paying for more views if you're just trying to crew reach. But what makes us unique and visible measures is that we're sitting on seven years of historical data and seasonal data too. So we can actually go back and if your goal is specifically, I want to dominate the, let's say it's August, we're gonna back to school timeframe. We can actually go back and look at what was going on in August of 2015, 2014, 2013. Look at which brands were the most successful in garnering viewership, garnering social interactions, garnering conversation, looking into what they were doing from a creative appeal, from a creative execution standpoint, how many videos they were promoting, at what frequency. And these are all sort of like very qualitative guidelines that we can provide to our clients on top of then running their media through our system and targeting who they want to target, hitting, optimizing for their goals and et cetera. So, and I know you can, we could parse people, customers' objectives and there's all kinds of segmentation you can do there, but I've found that it boils down to two things. They want brand amplification or they want leads that convert. Yep. Is that fair, first of all? Definitely is fair. And what's working in those two categories? So, in terms of amplification, something that we're really trying to provide to our clients is something we call a share of attention and that's really, we're trying to go away from looking at, okay, this is what your brand is doing in a vacuum and really looking at, this is what your brand is doing compared to your top competitors, compared to your space. And that's what's important because you can be delivering 5 million views per campaign and think that's fine, but if your competitors are actually doing 20 million plus then you're actually getting lost in the noise completely. So what we sort of provide is more of a holistic snapshot of that. Attribution is always a bit more difficult to sort of link individual buying actions with campaigns. We do partner with companies that help us do brand lift surveys to see whether or not, compared to a controlled group and a group exposed to a video campaign, whether or not a campaign is effective in driving lift, driving, purchase intent, and stuff like that. And your focus is on the measurement, not on the conversion or the amplification technique, right? You're there, I would imagine, if you can't measure it, you can't improve it. So that's kind of your business, right? Is that correct? That's correct. A platform would do this, or it's a service, you do it as a project, how does it work? So historically, all of this data that was sort of flowing through our system was sort of just staying within our system. And whenever a customer would reach out or sales out, we would do a question, it was definitely managed from a managed service perspective. But over the last year, we have built out a front end that sort of queries on demand the various reports that we've been providing by and to our clients, and this might include some of the competitive reports that I've mentioned. So we are developing a platform that automates a lot of this work. And what are you doing with Vertica? Autonomy or Haven? Absolutely, so from the active measurement perspective, so the stuff that we do run, we're getting about a couple hundred million new rows of data every day that comes in, that are associated with the views that we deliver. From a passive side, we're definitely handling much less daily volume, but we're also handling years worth of historical data from hundreds of millions of videos that we've discovered and tracked associated with brands. And our platform does some pretty complex queries and pulling metadata like which celebrities were involved in some of these campaigns, over which dates, over which platforms. And we found that Vertica was really the only solution that could handle the scale and the complexity of some of the queries we wanted to do at a time that was acceptable for us. When there's a big controversy in the publishing field about ad blockers right now, is that an issue for your business? How big a problem is it? Yes, ad blocker is definitely a big elephant in the room when it comes to ad tech. I will say that we personally have chosen not to tackle that. We sort of assume, so a lot of our custom ad formats are very user action oriented, so you have to click to hit play, you have to click to skip, or you have to click to choose which video you want to watch, which ad you want to watch. So our philosophy is that if the user has ad block going, they really don't want to watch ads. So we're not going to force, like we actually don't really do too much traditional pre-roll, because we don't want to force the user to watch ads when they don't want to. We want them to sort of opt into the experience, opt into the content, and an ad blocker is really just a filter for us to say, this person does not want to watch ads. Why? It's basically a wasted impression if we serve it to them. So you detect the ad blocker and then don't serve the ads, is that correct? Yeah, okay. And then you said, did I hear that right? You don't do like a little pre-roll or? So while we do technically have traditional pre-roll as one of our products, our core products are all based off of user generated views. So they have to click and buffer the video or they have to watch a certain portion of it and then choose to skip the rest of the ad. So it's a very user driven experience. And that's sort of fixed into our narrative that we're not really trying, we're not just there to take someone's video and throw it in front of another person's video. We're there to really help brands share their stories and their messaging across the publishers that we work with. Right, okay. And so you're basically trying to serve content that the users want to see. So a lot of that depends on the video itself, right? I mean, the video's got to be something of interest that can't be an ad in and of itself even though it sort of veiled as an ad, but it's got to be a value to the consumer. Yeah, absolutely. And so we actually do, we do get involved at times as early into the creative development process where we, because we have all these historical insights where we can look into, like I mentioned, the back to school time period. We can actually guide our clients in being this is the type of creative appeal that really does well in terms of gardening, viewership and conversation and positive sentiment analysis. So yeah, it definitely is a huge factor in the success of a campaign. Great, all right, we'll have to leave it there. Evan, thanks very much for coming on the queue. It was really a pleasure meeting you. Great, thank you. All right, keep it right there. Everybody will be back with our next guest. This is theCUBE, we're live from Boston. Right back.