Evolution in Mobile Advertising - Targeting, Creative & Media Buying

Editor's Notes

• #2: A year ago I wouldn’t be standing here.The reason?Because FB had no mobile ad business whatsoever,
• #3: And yet today, in under a year’s time, They’ve become the #1 publisher of mobile display ads
• #4: What represented 0% of the company’s ad revenues a year ago, Is on track to be $1 billion dollar business this year in the US alone.
• #5: intro
• #6: It seems like every year for the last five years we’ve said “this year is the year of mobile advertising”We’ve become accustomed to an incredibly slow pace of innovation.
• #7: But it didn’t take FB 5 years or even 2 years to come up with their mobile advertising solution.It took them a matter of months.
• #8: Just couple weeks after FB IPO’ed, it launched its first mobile ad product broadly.3 months later, mobile ad revenues represented 14% of their business.And it felt like every other week they released new mobile ad products, like the one designed for app developers pictured here.By year end mobile represented 23% of their business.And we learned last night during FB’s Q3 earnings call that mobile has now grown to represent 30% of FB’s ad business. So, why did mobile ad dollars move so quickly to FB?
• #9: Well, we can say with authority as one of FB’s largest development partners,It’s because the company is constantly innovating and iteratingTo solve very real challenges on behalf of marketers
• #10: And really the success lies in taking rich targeting data and compelling creative that has proved incredibly effective for advertisers ondesktop -- to mobile.Where in fact they’re see much higher CTRs
• #11: But there’s a second evolution happening in mobile.It’s also coming from desktop, And this time is being demanded by marketersIt’s about how we measure and buy customers
• #12: More specifically, it’s about the ability to measure and buy media for ROI.
• #13: So, today we’re going to dive into these two driving forces of change in mobile advertising:Facebook and ROI.Thanks to Facebook,It’s no longer acceptable to use light, inferred targeting on mobileNor is it acceptable to use paltry creative on mobileAnd thanks to marketer demands,It’s no longer acceptable to use proxy measurement and buying techniques that mask true ROI.
• #15: So, to start, let’s talk about the evolution FB’s driving in mobile targeting.
• #16: FB has brought a 750mm person audience to mobile.But I’m not here to talk about how this scale is double that of which most mobile ad networks offer.Nor am I here to talk about how engaged those 680 million people are – spending on average 15 mins a day on FB mobile, which actually exceeds the average time they spend each day on FB desktop.
• #17: Instead, I’m here to talk about the fact that each of those 750 million people have identities and an incredibly rich profile of targetable attributes.
• #18: No secret that fb knows a lot about us
• #19: Alternative, prior to FB: targeting on mobile lacks identity completely. Targeting is either not personal (eg device – iphone 5 vs 3gs sure is indicative of purchasing power, but not what theyd be interested in purchasing) or inferred (gender – pub aggregations, could be female, could be male)
• #20: And so this has resulted in an industry that relies on “spray and prey” tacticsIn other words, they’re buying blindly, while thinking opportunistically.
• #21: What’s funny is that the people in the room next door are talking about how to make these historical mobile targeting attributesmore “precise” – When you chose to sit in this room, because you know
• #22: That FB has clearly set the bar already, leaving the rest of the ecosystem behind.Marketers have an absolute trove of targeting data and combinations to leverage with FB mobile.
• #23: While I won’t go into all, I will talk about one – Custom Audience targeting.One we’re most excited about recently is that last box – CRM. With a targeting product called custom audiences, advertisers can upload information from their CRM database – emails, phone numbers – and Facebook will match this data with profiles containing the same information. You can remarket w/o needing cookies, or expand this to lookalike audiences w/ similar attriubutes
• #24: .
• #25: Here’s what we’re used to thinking of when we think of mobile ads
• #26: Enter FacebookIt’s pretty clearhow they’ve changed our expectations around mobile ad creative.They’ve brought ads to mobile that take up almost the entire screen This provides creatives a huge canvas
• #27: They’ve brought ads to mobile that act like content and feature social engagement from your friends –you actually want to engage with them
• #28: And finally, they’ve brought ads to mobile that are designed to drive specific behavior –Walk through each
• #29: Not the tiny ads on the RHS you think of when you think of FB adsCurrently representing about 1/3rd of the ad spend through our platform across desktop and mobile. Provide consistency across platforms
• #30: These currently represent ~25% of all the spend through our platform on a given day.Retailers, mobile gaming companies and travel companies have been quick to adopt these ad units. And it’s easy to see why:FB designed these ads specifically to help app developers to drive discovery and distribution of their mobile apps.1% ctrsInstall from w/i
• #32: In the beginning, the industry first started measuring success simply by way of view.The Cost per Impression metric was the gold standard for pricing.
• #33: Then tracking evolved and we were able to measure and thus pay only when ad placements generated a click.CPC became the defacto way to buy.
• #34: More recently this evolved into paying only when a mobile ad drove a specific action – like an app install.Leaps and bounds better than buying eyeballs and clicks, but not perfect –
• #35: Why?It doesn’t consider what happens after a person takes an action.And thus doesn’t consider the quality of the customer acquired.Did the person install and never open the app? Or did they come back and purchase over and over again? Who exactly did you just pay for?
• #36: Why?It doesn’t consider what happens after a person takes an action.And thus doesn’t consider the quality of the customer acquired.Did the person install and never open the app? Or did they come back and purchase over and over again? Who exactly did you just pay for?
• #37: Before we dive into an example, let’s ensure we all understand CPA, LTV and ROI.
• #38: Before we dive into an example, let’s ensure we all understand CPA, LTV and ROI.
• #39: Before we dive into an example, let’s ensure we all understand CPA, LTV and ROI.
• #40: Before we dive into an example, let’s ensure we all understand CPA, LTV and ROI.
• #41: Here’s an example.Say you’re a retailer advertising a tshirt.
• #42: And you’ve determined through taking an average of your customer acquisition costs,that you’re only willing to pay up to $10 to acquire a customer through ads.
• #43: Scumbag Steve here purchases a t-shirt, but never purchases again from the retailer.
• #44: While Mother Mary here buys the t-shirt, and in the coming weeks continues to purchase for her whole family.However, because of your $10 CPA -- you never end up actually acquiring Mother Mary
• #45: Why? Given there’s a bit more demand for her, Mother Mary has a CPA slightly above $12. So that leaves you with a customer pool filled with Scumbag Steves – who buy once and never return –and you miss out on your most valuable customers who continue to purchase again and again.
• #46: A 33% more expensive audience ended up generating 400% higher lifetime value and 28X the ROI.
• #47: A 33% more expensive audience ended up generating 400% higher lifetime value and 28X the ROI.
• #48: A 33% more expensive audience ended up generating 400% higher lifetime value and 28X the ROI.
• #50: Since you’re buying on a CPA basis, the ad network you’re buying from is actually incentivized to buy cheaper audiences than Scumbag Steve so they can pocket higher margins.Like Broke Brian.And Debtor Dan.
• #51: Saw this a ton in the last few years w/ marketers buying likes.“we want 1m Likes for $2 each”Ad firm: “sure” -- > goes gets $0.50 from MX, pockets $1.50And now, these companies are asking what they actually bought.Luxury brand example – affinity analysis shows incomes under $20k; aspirational buyers
• #52: Clearly there’s inefficiencies and dirty little secrets w CPA despite the fact that the industry has embraced it
• #53: And as we saw with the example before,since you’re not buying based on more downstream behavior, you’re buying opportunistically with hopes the action takers will turn into lifelong customers
• #54: And also as illustrated in the example, it offers a breeding ground for arbitrage
• #55: The average is actually discouraging buying high value audiences who may cost slightly more
• #56: In that same way it encourages buying low value audiences
• #57: And because CPA is an average,It values all potential customers as equalWhen some are more valuable than others.
• #58: Because CPA is an average and you’re not willing to pay more, you’ve limited the potential universe of people you can reach.
• #59: This is especially the case in times of high demand when costs riseLike the holidays, when people are buying the most
• #60: Last but not least,Since all your competitors are buying on CPA,All this demand is actually driving up the prices for cheap audiences!
• #61: So, how should we buy to avoid all those pitfills?A year or so ago at nanigans, after building tech that measured downstream,We realized highest ROI users are also the highest costAnd this is when we doubled down on our roi based bidding and optimization algorithms
• #62: Does this bidding strategy work? Let’s look at a few case studies
• #63: Goal: Drive new members in a ROI positive manner.After 6 months: 4.5 million member base
• #64: 38% of which were acquired through FB ads, impressive enough if you’re sitting in the audience thinking “fb ads don’t work” “no one clicks fb ads”
• #65: More impressive is the customers acquired thru roi based bidding were 400% more valuable.
• #66: So the beginning of this year we brought these ROI based bidding algorithms from desktop to mobileAnd the early performance has been really promising…
• #67: Acquire customers who not only install, but who purchase
• #68: Specifically, reach audiences that deliver a 49% return on ad spend within 7 days of being served an ad (the overall ROI it was achieving through other ad networks)
• #69: Incorporated our code snippet.Measured, attributed and optimized spend on an audience level to find the most profitable customers.In other words, our system allocated more ad spend behind audiences who were performing form an ROI perspective, and dropped spend from audiences who may only be clicking, but not monetizingCreativeMobile app install adsTargetingGender, age range, geographyBidding & OptimizationLifetime ROI
• #71: All of this illustrates that CPA, by definition
• #72: Means youre not buying for roi
• #73: So, evolution.There’s no doubt there has been a huge evolution in mobile advertising.
• #74: There’s been an evolution in what we expect from targetingThere’s been an evolution in what we expect from creativeAnd there’s been an evolution in how we think about buying media
• #75: Targeting is no longer inferred or non-personal
• #76: Instead, we have a vast and rich targeting set
• #77: No longer have only tiny ugly banner ads
• #78: Huge canvas that acts like content
• #79: No longer based on proxy metrics like cpa
• #80: Instead can be transparent into true roi
• #81: And all of this happened in under a year.And we can say w/ authority as fb’s largest development partnerThat fb will continue to build and innovate in mobile
• #82: And if you don’t believe us, I hope you believe him.Said over and over that they are mobilefirst company; not just to be trendy, but bc they realize the value they deliver to investors will be based on their ability to monetize mobile.Its where they’re placing their engineering resouresIts where they’re placing their partner resourcesSo now,just like you can’t buy ads on desktop without buying on FB,You certainly can’t think about buying mobile ads without buying FB
• #83: So you can se