Insights Learned From Our Roundtable
We recently recorded one of the best roundtable discussions we’ve ever had, hosted by incremental and Singular.
Leading up to, and with the launch of iOS 14.5, our industry continues to undergo massive changes. We asked Jayne Peressini from E.A. Games and Alex Weinstein, former head of Growth and Marketing at GrubHub, to join us and talk about these tectonic shifts.
We spoke about the opening rate for ATT, the profile of a marketer in this new world, implications to performance marketing, what you can do to prepare, and our predictions about if, and when, Google will take similar action.
The roundtable participants were:
It’s not the market that is changing. It's just the measurability conundrum. There were some theories that everyone will shift their budgets to Android, but I simply do not see this happening.
There's a different dynamic with partners like DSPs, where the decision engine is now being in-housed within our company, rather than it being managed by a DSP.
We have to do a lot of the legwork that used to be done for us by the tech ecosystem. It was quite a luxury.
I’m also seeing a potential rise in the importance of 3rd party app store. Zynga is doing some interesting things in this area. That’s all I can say at this point…
I've always been quite cynical on the topic of trusting anybody in the ecosystem with any of the numbers, because I'm a big believer that the incentives are massively misaligned. I tend to be of the view that you've got to do all the incrementality studies yourself and you cannot trust any of the media vendor participants, most definitely not the platforms or not the people who are doing the buying for you, because all of them are making money from you - buying from them.
Apple is now broadening the definition to all sorts of tracking. We're seeing the trends happening in the Web with third party cookies. There's a clear shift towards privacy that's just eliminating a certain set of technologies and solutions we had before. The changes are massive.
No one knows yet what the opt in rates will be, but probably, they will be low.
Will companies leverage the data they do get consent for and apply to the rest? I think that's going to be a really interesting exercise. I still think if you can get consent, it's important because you will have leverage to do other things like modeling, etc.
There is a reason why we named our roundtable: market like a marketer. In my view, what this change does is forces on the industry to market like a marketer.
When we think of old school marketing, we didn't have such access to data. We didn't have the ability to understand on a on a click or an impression basis who was the last to touch a user.
Testing meant – TESTING. Trying different mediums, different channels and so on. I think this is of the biggest changes that are happening on the macro level.
User opt in will just be another data point. Marketers coming with an offline background or agency background are actually a bit more comfortable in this new situation.
Apple is like the Israeli Mossad. They don't give any information before doing anything. That's the challenge with any business that rides on someone else's platform.
Mobile Marketing became synonymous with “growth”, but in reality, mobile marketing groups have shifted to becoming experts not just in user acquisition, but cross-selling and up selling. Athleticism in marketing becomes really important. There's a lot more interest to hire folks that have some creative background and can think beyond just linearly of “hitting a number”.
These changes are pressure testing the types of people that succeed in this industry, so I'm happy with diversifying the talent pool and specialties.
Marketers suffered from “Stockholm syndrome” for a while. They were completely in love with some platforms that made their lives too easy, but the challenge was that they could not recreate the success of those platforms. Marketing becomes harder and the profile of a marketer probably should always have been someone that can figure out and solve problems by trying out new methods, new ways of doing things.
We told this to Apple in the past – this move forces companies to have a really superb data science team to maximize the value of their measurement. If in the past you could just turn on Facebook and got superb results – you’ll need to work harder and make your marketing operations more sophisticated, or use more advanced methodologies like incrementality testing or media modeling. It does give an advantage to bigger companies as they have more resources.
I agree with what was said, but challenge the aspect that old school measurement is difficult. I would argue that going back to the fundamentals to the basic instrumentalities is actually not difficult. It's just that platform like Facebook and Google have taught us that Marketing meant looking into their fancy dashboards and trusting their graphs, always going up into the right. Platforms have been consistently spoon-feeding marketers, telling us that this is what is means to be “data driven.” In my mind, TV and digital video are just as much UA as UA is UA.
I agree, as much as we want to focus on the digital channels, connected TV, traditional TV, radio, and podcast are valid channels for app distribution.
Maybe we’re not getting real time conversion data, but with incrementally testing, we can measure value.
If you’ve not been pigeonholed to just digital channels, you might be in a better position to absorb the fact and the reality that the market's changing. You probably have some foundational methodologies you can work with.
We are at a moment where a broad set of marketers will be forced to learn about things like market incrementally. I really dream of a day when you say something like this in an audience of one hundred people at a conference, most of them will actually know what it means. I don't think an average marketer currently understands what it takes to do an on/off test…
The only time marketers were exposed to incrementality was what the platforms had to offer them. Whatever results they got would have always been either wrong or biased.
Incrementality was perceived as a dirty word. As witchcraft.
I think we've regressed somewhat. Marketers before the times of digital, had to figure out if the billboards around the city are actually making an impact on sales. The problem that we're currently looking at is solvable using those same means, but digital advertising forced a weird complacency to trust the ones who we shouldn't trust - the platforms that make money off of us (as marketers).
Many marketers experienced the on-off test when their budget ended towards the end of the month. But I know you don't see too many marketers going out to the press and saying, well, I actually found out that 80 percent of my budget was redundant, like Uber or Airbnb found out. No one wants to say that much of their ad spend cannibalizes their organic sales.
Incentives play a huge part here. A lot of marketers’ self-worth comes from the amount of budget they manage. This is how we’ve been “brought up”.
My team is not incentivized by how much they spend. And everyone is playing on the same playing field. When growth marketers are incentivized with how much they spend, they get away from the ultimate business goals.
I've seen decisions made that negatively impacted the business because of the incentivization being misaligned or being just purely based upon how much you spent and how much you drove.
We've seen the dynamic a lot when it came to Fraud and the fact that people didn't want to acknowledge it, or it wasn't always a priority. It would scare me to know that the same incentives still exist today..
Changing behavior because you feel like you've been defrauded feels really, really bad. But there's also an in-between, which isn't fraud. But at the same time, is not incremental, which I think is the vast majority of the wasted spending. Is Facebook actively defrauding you by having their algorithms select the people who are going to buy your product, just a couple of hours before they actually buy? No, that's their incentives and they coded that algorithm to be that great. They didn't design for it to be bad, but it is.
No one wants to save the waste. Everybody just wants to get the full value for what their budget is.
Decision engines are going to be imported internally, media trading, cohort measurement, etc. The day to day and the optimization changes will be what you need to empower your team to do, based on the methodology your company set, and how you arrive at what's contributing, what's incremental, what's not.
This change is making the average marketer care about the overall business more. And I would get that if we fast forward three, five years out of an average market, it will be much better able to answer questions like: “How do you know that this is working?”.
Rather than mastering setting up campaigns, marketers should master experimentation because that will be the only way to survive.
And right now, marketers need to just trust whatever they see on the networks and hope that the networks give you the full visibility.
I do think that Singular was and IS in a fantastic position to visualize and be the neutral auditor for attribution postbacks, even if it’s coming from Apple.
Singular always stayed away from bets on aggressive sort of device graphs or using a lot of third-party data. The tech stack moving forward is going to be more complex - while we need to support new things – all the current tech (for IDFA attribution and tracking) is still going on in parallel.
There’s a lot of talk about media mix modeling. But MMM is not really an alternative measurement method. First, you do need a ton of data, the more the merrier. 12 months usually is the minimum. You need external data regardless of if it's weather, politics or economics, or anything else that might have affected your sales. All marketing activities, competitor information and so on.
Using all of this data, you're trying to triangulate what worked on a macro level, it's really good when you are running on multiple mediums, but rather useless for digital only advertisers.
What is important to remember is that it’s never “MMM OR Last Touch OR Incrementality”. It’s not this vs. this, but it’s This + This, depending on the use case.
MMM is more of a planning tool. I wouldn't expect my team to use MMM for daily optimization. For me, it was easier to just pick A methodology and move on to other business problems.
I'm a big non-believer in the context of the concept of media mix modeling at all. I don't think it's actionable. I don't think that it can be good irrespective of the amount of marketing spend that you actually put in for anything but the most transactional of the businesses. I'm a believer in doing two things. One is incrementally testing to understand the impact of your additional investments in individual channels in the context of the rest being held constant, and the second is tactical measurement to try to make the campaigns that you have and the channels that you use more effective within the specific context.
Real-Time attribution remains tactically important. If you are spending a lot of money on, and the Red Banner works better than the Blue Banner, you don't need any sophistication other than last-touch attribution to tell you that the Red banner works better.
Last click was really the predominant method, partly because it was the only signal large platforms like Facebook provided.
Multi-Touch attribution was like showing a banana on a slide. There was never any consensus on the approach and how much weight each signal should get.
Last-Touch has its faults – but at least everyone can agree on it.
Start caring about what the CEO is caring about, taking a higher vantage point. I think it will cause the right questions to pop up.
I don't think you should be in growth leadership if you don't align with your CEO on business goals. A growth marketer not aligned with their company would be a very dangerous person in the team.
Obviously, if you have not yet figured out a conversion value, you should.
My tip would be – explain to your company “what changed”. There's been a lot of changes and I've seen that it's overwhelming people. Help your company through the transition, teach them the new terminology.
I think they will remove the GAID but keep the referrer mechanism. This will allow them to raise the “privacy flag”, while not disrupting their own eco-system as much.
I've done a historical chronology recently about identifiers and noticed the same pattern - Apple does something. Google does something. Apple does something. Google does something.
In my opinion, Google will announce the deprecation of GAID in summer (2021) and execute end of 2022, giving the market, especially their own tech, time to adjust.
I think differently. Google will only make a move as a monopolistic move, damaging smaller publishers in the process. I think that their brand does not stand for privacy. I think they will do the absolute bare minimum. I think if they are not forced by the rest of the world to make this move towards privacy, they will take no move at all. And if they are, they will do some small step in that direction.
Google will surely be more considerate of their ad business, but they will likely be pushed into a privacy centric move as well.
I think that the message of “privacy” sells devices, which is Apple’s interest, but not Google’s.
The reality is that WE as proud members of our eco-system never abused identifiable data. We never even had access to understand what our friends are doing, or what someone’s internet browsing history looks like – but these “scare tactics” of Apple saying: “our devices are more private” sell devices. So what if they also empower their own ad business with new placements that don’t play by the same attribution rules? Business is business.
Singular is the leader in unified marketing data and analytics. Marketing, BI and Analytics teams from Lyft, Airbnb, LinkedIn, Rovio, Stitch Fix, VICE, Warby Parker, Acorns, Twitter and more choose Singular to unify their siloed data, apply attribution, and expose insights to make faster decisions. With thousands of integrations, easy set-up, accurate and flexible analytics, plus automatic loading directly into your BI tools, we're the one solution your team needs to gain a complete and real-time view into marketing performance. Visit www.singular.net to learn more!
INCRMNTAL is a continuous incrementality testing software, allowing marketers to understand the true incremental value of their marketing activities. The platform points out activities adding value, activities that take value from other activities, as well as activities that add no value. The technology behind the platform works using Causal Inference, using attribution data from MMPs or SKAdnetwork in aggregate form. The platform is completely self-service and requires no SDK or code for integration. Visit: www.INCRMNTAL.com to learn more