[fallous]

I've been cogitating on some of the themes Standage brings up in this interview for awhile and I suspect he's far more correct than incorrect. The proliferation of adblocking and incognito browsing is going to do marked damage to advertising networks like DoubleClick because the loss of demographic targeting means they are forced to basically become run-of-site ad outlets. Additionally, the amount of fraud in the ad views and click rates has become far more obvious to advertisers over the last 18 months and will only increase as sites try and game the revenue to make up for ad blocking. This does not bode well for the traditional ad model that has prevailed on the Net for the last 15+ years.

I suspect that Facebook becomes the real winner in this due to the fact that no one uses the site without logging in, has provided enormous amounts of targeting info, and with Facebook controlling the ad experience they cut out the likes of Doubleclick and deal directly with the advertisers. They're vouching for the legitimacy of the ad performance metrics and handling the targeting internally using their own data, and so will be able to command a huge premium. A few years ago I would've laughed at the idea that Facebook would be able to compete with Google in the ad revenue market, but I'm not laughing now.

[enraged_camel]

Don't adblockers affect Facebook as well?

As a Facebook user, companies are spending money to show me targeted ads, but I never see them. No doubt this decreases their ROI.

[dragonwriter]

> Don't adblockers affect Facebook as well?

Perhaps, but a major part of Facebook's business model is organic advertisement (pay-for-reach posts) which won't be affected by ad blockers. AFAICT as a user, that's an increasingly important part of Facebook.

[shostack]

While you make some good points, I think you may be underestimating the feedback mechanisms of the industry.

Viewability measurement is in its infancy, and still has many issues. That said, it is fast becoming a standard measure, and any savvy display advertiser will be using it. At the end of the day, the ones most at risk are brand advertisers who aren't running direct response campaigns. They may have ad recall studies and Marketing Mix Modeling setups that give them some visibility, but there will always be some inventory wasted at the scale those advertisers are dealing with.

For direct response advertisers, things are getting better. While it may be harder to target some segments, and that % may be growing, the tools advertisers have available to them simply did not exist even a few years ago.

Take for example Google Analytics attribution and multi-channel reports. They have made a basic cross-channel attribution toolset available. To everyone. For free. Think about where the industry was even five years ago and let the impact of that sink in.

Facebook's acquisition of Atlas, Google's acquisition of Adometry, and other recent acquisitions in the bid management/dedicated attribution platform space are all pointing in one direction: better visibility into the contribution of your individual efforts.

I manage digital advertising for a living and have done so both client-side and at a top agency in the space. When it comes to the space, and in my not-so-humble opinion, the #1 thing clients wanted (and that I want doubly so now that I'm client-side) is better attribution.

When you are sitting on the types of data FB and Google are, you have many pieces of that puzzle, particularly the cross-channel and cross-device pieces. You also have the statistician brainpower and engineering talent to create the modeling tools that can give this visibility to advertisers. Why they haven't pushed harder on making that data visible is anyone's guess (I do wonder about how any negative revelations might impact their business), but the acquisitions seem like a big step in the right direction.

That said, digital media needs to go beyond static attribution models like: linear decay, time decay, U-shaped, first touch, last touch, etc. Instead, it needs to move more to the dynamic models, where data is assessed at the individual conversion path level. When you look at attribution at the channel or even campaign level, you are missing a ton of the story since the impact of say, a generic video ad vs. a laser-targeted retargeting ad can be night and day. Sure you might want to see channel data in aggregate, but you can't effectively optimize much at that level.

The advertising bubble savvy individuals in this industry are aware of those is video advertising. There's been a big hype train, CPMs are frothy, and everyone is switching to some sort of auto-play/auto-play-next-video format. Personally, I'm not convinced the value is really there at many of these CPMs, but I test and let the data decide.

The best thing companies like Google and Facebook can do to safeguard against any growing mistrust of online advertising efficacy is to keep improving their attribution tools, particularly for display and view-through performance.

Knowing what I do about the data available to me and its strengths/weaknesses I sometimes wish I was back in the days of last-touch models. There's few things as painful for me as knowing that better data exists to optimize against, but not having all the tools I need to get it because they are prohibitively expensive still for many budgets.

[blumkvist]

You expect google and facebook to do attribution modeling for you? That's simply unreasonable to expect. That's like waiting lead scoring in CRM systems to be accurate. Never going to happen. All they can do is put some APIs in place, so you can get a hold of the data, but the modeling part is up to you.

[shostack]

I expect them to make their best attempt, and give me all of the tools/data they have available to the best of their ability. More to the point, I think it is in their best financial interest to do so.

There is a lot of doubt around the contribution of display, and video as well. By helping arm advertisers with more data/tools to help them understand the impact of the efforts (and optimize against that), they further prove the value of their channel. This would likely help grow revenue as the value becomes better understood. Depending on what that value ends up being, it may even grow margins if they can command higher CPMs as a result.

I've seen you post before, so I know you have a solid understanding of the limitations of the data currently available. I want to be clear that I'm not asking them for a silver bullet. There will always be a need to analyze performance, view the data through the lenses of different models, and apply your own judgement.

But there is most certainly some low hanging fruit that would be a step in the right direction. Let's take the AdWords platform (not using any bid management platform) for example. Off the top of my head some improvements Google could make there would be:

- Making view-through revenue data available within the UI. If I have a conversion tag placed and am passing in revenue, Google has this data. Why they can't display it when they can display VTs is still a mystery to me.

- Search Funnel reports are a nice start with their attribution and path analysis tools. However the fact that they don't have display campaigns in these reports is a real limited. I know you can now get a lot of that in GA, but many advertisers don't have that setup properly, and doing anything at scale in GA these days is a PITA with their reduced sampling threshold.

- Google has the path data, yet for some reason it is still not exposed in the main UI outside of "assisted" metrics. Good luck digging into that data at granular levels as well if you have multiple conversion goals. Custom columns are a start, but they don't go down to the keyword/ad level yet and are still quite limited.

- Google pushes video ads hard, yet the Video Advertising product is for all intents and purposes a separate product. Sure it shares a similar design, but most of your data is siloed from AdWords (with the exception of audience lists).

I could go on, but these highlight some areas where they could do more to prove the value of their offering. When you throw the capabilities of a platform like Adometry into the mix, I'll bet they could throw a LOT more meaningful data at advertisers. Why shouldn't I, as a customer spending large sums of money with them every month, expect access to these capabilities to get the most out of their product?

I realize your point was that they can provide some data, but decisions still fall on me. That is valid, but there's still a lot more they could do to help inform those decisions with what they have available.