Connect MORE Demand, Make LESS Money | PGC Digital #1

hi everybody my name is Dave Simon like like you mentioned I'm the CRO at fiber we're an app monetization platform we live and die in the App world we have a very heavy emphasis on gaming in particular but by way of background I actually come from the programmatic space I've only been in the in the in-app space at fiber for about six months prior to that I spent the last 15 years working in programmatic and having come to this world where programmatic is still very much new and a sort of a new thing that it's being tested I thought it'd be worthwhile to share a bunch of the learnings that I've gained over the years from when web publishers made the transition to ad network driven monetization and into this notion of exchange bidded based marketplaces and so I figured I'd go through some data that I've already seen that we're using to help our customers make smarter

choices about how they are monetizing in the land of this new programmatic opportunity so really quick about fiber we're we're a global company we're publicly traded on the Frankfurt exchange so our corporate headquarters are in Berlin but we've got offices all over the place we're 250 employees we're 50 percent R&D and product a hundred percent app focused and you know it's one of the nice things about about being a company in this space is that because we're public you can see where we are you can you can understand you know there are certain standards that we have to adhere to around transparency and around how we behave that hopefully gives a lot of our customers some confidence in in the kind of company that we are having spent the last 15 years in and out of public companies and private companies I can tell you when you have to file things in a public report there's just a certain amount of transparency you you're

willing to provide so so pretty pretty helpful when in the context of the app space which is still very a nascent area okay before we talk about but before we talk about how the mechanics of this work I wanted to share with you guys some research that I've done using apps fly or emarketer data about how big is programmatic in the in-app space it's a question that I've asked all the time and the undertone of that question is why do I care about this right I'm making money I'm getting paid like do I really have to sort out this programmatic theme the in-app global ad spend market for just in-app almost all of which is app install based is a seventy six billion dollar market that's growing to almost 120 billion by 2022 so think about it it's a market that's growing at about 50% in a global programmatic market that where growth is actually slowing down so if you look at the total global programmatic market of

about 200 billion that growth rate is gonna go from 35 percent to 20 percent over the next three years the in-app space is expecting 50 percent growth in programmatic and so I'm sorry the global ads been in in in-app particularly in particular programmatic today is 47 billion dollars of that market so a huge percentage of the total amount of money that's spent in app is spent programmatically now hopefully these numbers make it clear to you that this is an area that if you can unpack the value in this supply chain of programmatic there's tremendous upside to be unlocked ok so this this is the market itself there's an analogy I'm a big fan of using for how we all enter the programmatic market as sellers in our minds as publishers we look at our waterfall and we say okay I have these 10 partners with these 10 tears and

their fill rate is X and their eCPM is y and so this is a very organized thing in other words we think what we put out into the market to the advertisers who want to buy our supply looks like a well organized Whole Foods where you know on you on avocados you go to the avocado stack you want a watermelon you go to the water mode you know exactly what to get and how to get it in reality when we publish our available inventory into the marketplace for buyers to buy it looks more like this where there is no rhyme or reason for how things are priced why they why they're able to win why you're able to get certain inventory at a certain price it's chaos and it's incredibly hard for the buy-side to figure out how to navigate this chaos if you think about it from the perspective of a you a marketer or any marketer for that matter buying ad inventory is a part of the process of selling something of driving revenue so what you're looking for is

consistent access to users at consistent prices that make it easy for you to forecast how much revenue you're going to make when we sell inventory programmatically today with the current construct we are making it incredibly hard for buyers to understand how to access supply at each of the price points that work for them in a consistent and reliable fashion so I'll walk you through why that happens and then I'll give you some data examples of the kinds of things that you guys as app developers and sellers of media of ad space should be thinking about when you're when you're listing your inventory this is a very very simple version of what we call a supply path from the time that an advertiser decides to input his budget to the time that the publisher gets paid this is the programmatic supply chain you put on a little opt out notice there oftentimes your vendor will do this for you then the budget gets put into a programmatic DSP a bidder

it points bid bids to an SSP like fiber fiber then renders renders the ad via the the mediation platform in this case I've used Mopa and then the ad shows up on the app on the apps on the apps inside of the app very very simple supply path if you were to take a look at how a network would work imagine eliminating these two steps and just having this the one section where a bidder and the SSP are conjoined as one and there's no transparency into who takes what rate or whatever you just you throw an AB dollar into a black box and the app publisher gets paid on the other end and and you're good to go very effective hard-to-understand not not very transparent in reality most app developers in fact every single actor I've ever talked to works with multiple partners and so rather than listing your inventory in one place you list it in

multiple places and those places go out and source demand oftentimes from the exact same place so to use real-world real-world partners here imagine that you're an app that uses multiple mediation you have a plugin in your waterfall you have fiber in your waterfall you have iron source in your waterfall all three of us are going to kick off an auction where we make that impression available to a third party bidder like a liftoff or a Moloko or a cross install and the result is Moloko or liftoff and cross install we'll see the exact same impression more than once often times at different prices if you extrapolate this out to a real-world example this is a large hyper casual publisher that we work with we mapped every supply path they have this is actually a relatively simple version of a real world example of how a publisher would be monetizing now these all aren't being called at the same time as this graphic makes it look they're all being called simultaneously based on on where

those those partners are in the waterfall but that doesn't matter because at the end of the day the bitter still sees sometimes 30 or 40 different opportunities with different prices to try and bin bid on an impression and so the result is a few things first the DSPs some DSPs don't know where to bid do you bid at the top do you bid at the bottom how much of the performance is reliant on the SDK partner that you're bidding into versus the inventory itself all sorts of unknowns arise when you create this scenario and by the way this is a relatively simple one I can I can point to examples where I've seen app developers who have 500 or 600 simultaneously available paths to a single impression and if you think about your own waterfall think about every partner at every price point as a supply path and then add them all up and you often we'll have at least a hundred and we

like I said we've seen up to 500 600 of them the other big advantage is that really savvy bidders will start to figure out which pads give you a better ability to win when all things are BR equal otherwise so for example in this in this scenario I have a $5.00 bid that I'm submitting through a pub Matic path it could be a SMAW towpath it could be an iron source doesn't matter a five-dollar path I also have a $2.00 path but for whatever reason my ability to win is actually better at the $2.00 price and it is at the five-dollar price there are a whole bunch of reasons for this sometimes it has to do with the technical connection between an SSP and the mediation platform how good is the SSP that you're bidding in to actually delivering a rewarded video ad unit what are their survival rates well in a lot of cases a $5.00 bid that you actually

win will never serve the ads so it's not effectively a served ad because they couldn't render it other times you'll see instances where a partner will have a $5.00 bid but they're sitting at a 1 dollar price tier and so the $5.00 bid can't win because it never actually makes it to the publishers mediation solution because you've set that publisher that that SSP at a $1 bid so all sorts of examples where lower bids will actually beat higher bids even though what you really want is to create the competition and give the market access to the supply as broadly as possible and in this particular scenario I absolutely shows one where fiber looked the best because you know why would I choose on one we look bad but there are lots of examples where fiber has a higher win rate at the same price tier than other SSPs full transparency there are examples where fiber has a lower win rate but again not going to put that in my deck I'll let a competitor do that instead so let me

give you some data of what happens independent of you know you stick fiber in your mediation platform at a $3 price what happens behind the scenes what's happening on the other end with me and my DSP partners that were bidding that are bidding into us okay hyper casual publisher very very large this is a 7-day snapshot of the number of bids we got at each price point now this publisher it had slotted us at a dollar twenty Tier four banner inventory so no matter what the bid I got back was I was paying this publisher a dollar twenty well you'll see here on the left on this y-axis is that my price is a dollar twenty a little over two million of the bids that I received came back in that price range between a dollar and a dollar forty nine pretty good right this is that the demand that I'm sourcing aligns pretty closely to the price I'm paying the publisher that's good

independent of that you see a whole bunch of other examples where publishers were receiving bids that were significantly higher than the dollar twenty tier that I was paying them at oftentimes north of two hundred and fifty dollars or remember I'm paying the dollar twenty this is a static price tier that has been set in the mediation platform at a dollar twenty four this banner placement for fiber so the seven day revenue the publisher got was a little over five K if you look at the potential revenue based on each of these tiers and the bids that I got back instead of paying the dollar twenty that I should that I did pay if I had paid the real market value there was another twenty seven hundred dollars worth of revenue potential of fifty three percent increase on one banner placement simply by accurately expressing the market demand and the market price for the placement that I'm currently only paying a dollar twenty

four so hopefully that makes sense it's just some questions afterwards um here's another example on this one's even more extreme this is a twelve dollar static price that we were set in the waterfall for an interstitial placement again a little over five K this is a different publisher this is another hyper hyper casual publisher again same thing we're in that twelve dollar price range we're seeing the majority of bids come back at the in that in that bucket where that's an accurate you know roughly close to the margin or the price that would be shooting for but there are still lots of bids that are coming back what's really a well above that price what's really interesting is we actually saw a whole bunch of bids we saw 2,000 bids that came in at $1500 now 2,000 bids doesn't sound like a lot but this represents six thousand dollars worth of revenue potential where somebody out there was trying so hard to buy this

inventory they looked at the inventory that this publisher was selling and said give it to me I'm willing to pay fifteen hundred dollars for it and the result is that publisher didn't win very much they only want about a thousand impressions and so what's really interesting here is this publisher probably doesn't want this buyer probably doesn't want to pay fifteen hundred dollars forever but there's somebody out there who knows that the value of this inventory is extremely high and they want to be able to access this inventory and so as a seller what you want to make sure you're doing is not locking this publisher in at a twelve dollar price behind some black box wall in the part form of a partner through fiber right that you really want to figure out what's with you who these buyers are and why they're willing to pay fifteen hundred because they might not be wanted to pay fifteen hundred forever but they might want to pay fifty and that's a 5x increase or a 4x increase over what you currently slotted this placement at same thing again here in this case we had a four

dollar and fifty cent placement bottom of the waterfall for this interstitial placement this is another casual gaming publisher four thousand dollars in a seven day period for this particular placement the interesting thing here is that the majority of bids came back well above that price this is a classic example of the publisher undervaluing the floor price that their inventory should be traded at if they had instead been able to to show the actual market impression they would have gained another [Music] sorry I missed a slide here another I think this one was another $4,000 it was like it was like another 120 percent increase over the over the existing inventory the existing yield same thing last one I want to make here this is a really a mid tier priced RV placement 7-day revenue was 11 Hunt was about 1200

bucks again more of the demand came back at significantly higher prices now the interesting thing is this placement in the art price in the waterfall is set through Auto CPM floor so the CPM will move but what you see is that the number of bids that we're getting back are significantly higher than the price that we've been quoted at and so again another 120 percent potential revenue increase if all you were able to do is accurately capture the bid price or at least understand who these buyers were okay last couple of takeaways and I'll wrap up so we can get to QA programmatic is a real thing in the app space if we look at what happened from 2008 to 2018 on the web you'll see that publishers who leaned into programmatic understood the supply chain demanded transparency from their partners and didn't just settle for the check are the ones that are unlocking huge budgets at low operational cost today if you think about the amount of money that we should I showed you that's left on the table

think about how your LTV calculations would chat would change if you knew that you were able to actually monetize certain users at the market value rather than an assumed average that's either said statically or dynamically via 24-hour API lag last thing I would say these are the questions I would leave you guys with to be asking of your partner's um how much of the demand that you're bringing to me comes from DSPs versus demand that you've sourced from a direct sale who are the demand sources that you're innovative with I want to know who the buyers are that are bringing the campaigns to my app via your platform and then most importantly what's the spread between the price I've set with you and the average bids that you're seeing if you understand those things you can start to calculate how to move your floors around how to rearm connect your waterfall and how to capitalize on the the opportunity to capture more budget so I'll pause there that was fast and furious hopefully that made sense and there are some questions right thank you

David thank you very much that was great very very very insightful and full of data and we've got a couple of questions in the question and answer window there if anyone else has anything they want to ask please do drop that in the box at the bottom there we've got we've got three there I'm going to just start and I'll read the questions out now in case people can't see them but you've got I've got a few there Roberto you've raised your hand it's probably yeah we are we aren't going to be able to unmute your mic so you might want to drop a question into the into the Q&A box so um so David we've got um first of all question from say there he says so do you provide the big price versus the number of bids graphs to add sellers yeah so this data that I showed here I would tell you who the publishers are but I didn't have a chance to share all this data with the pub's the short answer is yes that one of the critical things that fibres focused on is helping show the difference between prices that you're that you've asked to be paid and what the market is willing to pay and so this data here is available to fibre customers right now and I'll be honest

with you this is not unique it's not some custom report we've built every platform on the planet has access to this data and so regardless of who you're working with ask for this information if they aren't willing to share it that should tell you a lot because - the second question I think there was another question about it what's what happens to the spread when I get paid of doubt when I'm paying a dollar 20 and I'm getting $100 bid what happens well the first thing I'll point out is that $100 bid doesn't necessarily mean $100 paid right that that we do set we use primarily second price auctions so hundred our bid will be reduce and so if they're the only bid in there then you know there's since there's a there's a big drop and that bid - the clear price but even when we're searing clear prices like we saw here yeah the average network will keep some of that budget and what's interesting is that and a lot of networks are not willing to expose this data because it the amount of inventory the amount of

demand they source that comes directly from their own campaigns versus the demand that's the source from generally available sources that I tap into as well the amount of money that's being sent through DSPs is growing up like crazy and so this problem is getting worse not better and the only way to stop it is for publishers to try and claim that information back that's great we've actually got a bunch of questions now I'm going to skip any at one of our anonymous questioners as I see you if the coab in 19 simp acting-out medicalization yeah it's a great question I'm sure this is being asked in every session what I can tell you is if you live in the in app space we're primarily your inventing in particular gaming is in a wonderfully advantageous spot relative to two brands I'll give you guys an example though one of the biggest brand buyers that I had buying us in q1 is a restaurant chain called Olive Garden they're not spending any money right now and so if you're an app

who's 99% of your demand comes from other games probably very little impact I know there's a whole bunch of question about like user engagement rates and what happens to disposable income and I ap I don't have a good answer for that I don't think anybody does yet anybody who says they do needs to pump the brakes and see what happens because I've been looking at data from China we have a pretty big presence in China they're our numbers look pretty strong in China and and so like I I'm gonna hold off judgment so that becomes an answer great thank you I'm just very quickly want to answer Adams question about accessing recordings yeah you'll be able to watch this back later we're recording all the tracks so if you any late and miss what they were saying they'd run there'll be a recording made available later so don't again we've got a few more questions this to follow Stephan I lost is it right to assume that header bidding mediation would help solving this mishmash mismatch between demand and supply yeah so this is really interesting first I spent a year before

this at a consulting firm called John's meteor I have no financial involvement with that company but I would say that they are hands-down the best research firm out there for programmatic counts jou nce check out their research this is this is the kind of thing that what you'll see right now the inefficiency that we have in the in-app space is a result of a linear mediation platform so the price you set versus the price you're able to access our highly discrepant when you move to header bidding and you get away from that static set price and you let the market determine there are still big in efficiencies that can be exploited on the buy side and ways you can exploit it on the sell side and that's the 2.0 version of this so if we get together again and we're and we're seeing header bidding installed and it's at scale I'll give you another presentation on send some data around that because I think it's a great question but the short answer is like not yet header bidding will help but it won't solve all of it

we've got a few minutes left so well if you don't mind taking another couple of questions baby we've got some more in here another anonymous attendee asks I heard somewhere that Google and Facebook limit the line item calls is it true what are the ideal number of my Lighting's for Google Facebook and other popular network yeah so it's a great question and the answer varies widely that's not a helpful answer so I'll give you what data I have about Facebook and Google and and how they tie into this whole thing because I think it's important one of the main reasons way back up so fiber has a mediation platform called fair bid we talk a lot about header bidding the only way to really make this to make this problem get significantly better is to get away from the idea of line items right now fair bid was built to do that but a huge chunk of demand still sits in the hands of the ad networks so Google as a demand provider through the AdMob install stuff

they don't bid into any other other platform right now so Google's at Google doesn't give you access to their demand via a solid programmatic channel unless you're using AdMob right and so Facebook does Facebook bids through as a private programmatic mediated network into Austin to some others and so you can access Facebook demand programmatically when you do that you don't have to worry about line items anymore right because their bidding at every single impression into a real-time option and that option should reconcile against all of the other demand sources that are also competing and so the real quick answer to your question is the number of line items I want as few as possible and only where the partners are providing unique access to demand that I cannot access programmatically so the top of my waterfall I would put Google then I might put like a bungle right who

doesn't do any programmatic bidding but everybody else I would say you're gonna compete from my inventory inside of that auction that sits at the top of the fair bid mediation platform and by the way lots of people have auctions it's worth digging into why they're different there are some differences but by and large anybody that can bid shove them in the auction anybody that's big enough without access that you can't access programmatically stick them in a line item so for me I think maybe three tops or I should say three partners not three my items probably maybe not two minute or two more we've got three questions in there and I should say if anyone's asked a question or wants to ask a question or want to do it quietly they can probably find David through the meeting system after this if they want to approach you but we've got a question here just calling back to you're a coda 19 question another anonymous attendee is asked but they're concerned about working without networks which have net 45 or net 60 payouts during the COBE 90 situation what do you suggest yeah

okay so I'm gonna answer this question we sent out an email to all of our publishers about two weeks ago where we offered accelerated payment terms because we know for the app developer community you know we we don't want people sitting around waiting for cash to pay their developers we want you guys to continue to release titles so we've offered to accelerate to accelerate payment where we can wear again unique and that you know we're a public company we have healthy cash reserves we also have access to capital what I would say is that you want to be paying a lot of attention to the sequential liability terms I think there's going to be people at the very top of the ecosystem who will be continued to pay very very timely and a lot of people who are going to have a hard time paying because they're not getting paid on time so that like it's it's a it's a complicated thing what I what I would

say also though and I spent a bunch of time with bankers in the last two weeks there's gonna be a lot of liquidity you know being wine Mellon is a big bank here in New York they've got two hundred and fifty billion dollars in cash that they that's their cash reserve every day that's up a hundred percent in the last three weeks they averaged about a hundred billion in cash a day everybody's sidelined their cash which means everybody's gonna have access to capital what you want to be doing is getting paid by people who have viable businesses because those are the people that are going to have access to the credit lines and so if you have a long tail ad network that represents six percent of your total revenue and you see them spiking maybe pay attention if you're talking about an asset that's owned by KKR or Blackrock or that's a public company or something like that I wouldn't worry about it hopefully that helps