Does anyone believe ad products actually work?

Snapchat NRBM24
Jul 25, 2018 27 Comments

Here’s my take on digital advertising after 10+ years in the industry and having worked at Google, Snap, and in the agency world. I’ve personally spent hundreds of millions of dollars throughout my career.
————————————————————

-Attribution is still absolute dog crap. No one knows how to properly attribute conversions to media channels. I’ve worked with many of the top attribution platforms (both FB, Google, and independent). The data is barely actionable.

-Many ad agencies call push channels “brand marketing” and use awareness/consideration lift as measurement because they can’t prove value to the advertiser using CPA/ROAS. I’ve yet to see an agency or brand scale new demand efficiently using ads. CPAs almost always come out too high for advertiser to continue to invest.

-The marketing funnel as we know it is a bit of a joke. Most ad products aren’t designed to target individual users sequentially.

-Pull channels are always efficient with low CPAs and high ROI. I’ve run dozens of tests including less sophisticated (on/off and geo) and more sophisticated (ghosting). For every dollar you invest in pull channels you’re lucky to get a few cents of incremental revenue back - those users were likely coming to your site anyway. If you have decent SEO (and as long as the space isn’t overly competitive) pull channels barely contribute anything.

-Pull channels are likely just an additional cost to what you’re paying to drive the demand in the first place. I.e. generate lead through YouTube at an inefficient CPA and then retarget to same user at an efficient CPA - you just paid twice for same user.
———————————————————-

I watch agencies “make it work” all the time by blending ad spend. I.e. spend a little on push and most ad dollars on pull. Then give the client blended CPA/ROAS.

Push channels - YouTube, Twitter, Pinterest, Snap, Yahoo, DoubleClick, GDN, Google non-brand search, Facebook, Instagram, Gmail

Pull channels - Google brand search, Bing, GDN retargeting, Google Shopping, Amazon display (the most atrocious attribution out there), Amazon search ads

Agencies I’ve worked with through my career: OMD, AKQA, Fetch, Vayner, Essence, various small shops

Some of these channels are both push and pull, I realize this. This is my best effort to bucket them.

These are generalizations but hold true for 98% of advertisers with the exception of app install with high AOV and high margin/low price point impulse retail products.

ANYONE DISAGREE?

comments

Want to comment? LOG IN or SIGN UP
TOP 27 Comments
  • Tableau Pete@Nerv
    I’ve bought several things my friends already bought thanks to Facebook but generally the rest I never click on. My friends are apparently great at finding obscure things I also want. For example I didn’t know there was a power wheels type thing that looks like a Star Wars landspeeder.
    Jul 25, 2018 3
    • New / Product
      nicebot

      New Product

      BIO
      Smash that Private Message button
      nicebotmore
      That sounds awesome. Pic?
      Jul 25, 2018
    • Tableau Pete@Nerv
      There you go
      Jul 26, 2018
    • New / Product
      nicebot

      New Product

      BIO
      Smash that Private Message button
      nicebotmore
      Noice
      Jul 26, 2018
  • Microsoft / Eng uggamugga
    That's a lot of words
    Jul 25, 2018 0
  • Google AWASU7
    Ads can easily grow a business. Just watch shark tank to hear a bunch of success stories.

    You likely don't think they influence you but it does. When you shop for laundry detergent what brand did you choose...why that one. Likely you recognized the label due to advertisment
    Jul 25, 2018 2
    • Snapchat NRBM24
      OP
      These arguments are a bit anecdotal for my liking. I’m talking about concrete facts and controlled a/b testing.

      Here’s an anecdotal argument just for the sake of arguing. Brand awareness for Tide, Downy, etc is already at 98%. You already know what brand you’re going to buy before you walk in, you just don’t know it. More likely attributed to quality of product or that your family has been using it for years.
      Jul 25, 2018
    • Uber / Other
      Ehfink

      Uber Other

      PRE
      Uber, Google
      Ehfinkmore
      There’s lots of evidence to back AWASU7 Up. OP have a read of ‘how brands grow’ by Bryan sharp if you want some scientific evidence behind how certain types of advertising have worked well for traditional fmcg brands. He does actually use a scientific approach to measuring effectiveness of marketing on retail brands.
      Aug 27, 2018
  • Microsoft compiler
    Can you tell me how to learn more about this business? Eg CTR, avg bid prices, number of advertisers, benefits to advertisers, different countries etc.

    2 of the biggest employers are in this business, but I feel I don't understand it at all. In my 17 years of using internet, I have probably clicked on 2-3 ads by mistake. I mostly never even see them due to adblock, so i never understand how these companies keep reporting 10 billion of revenue every quarter.
    Jul 25, 2018 1
    • Walmart.com / Eng data_007
      Same here. But a lot of people do click on them. Especially, the ones on Google, FB, Amazon, Yelp etc. since they are so native. Its difficult to not click on ads on these websites unless you are explicitly trying to avoid them.
      Jul 26, 2018
  • Intel sillycon
    No disagreement ... But it's an amazing cash cow for some companies and the only reason they still exist and are profitably viable.
    Jul 25, 2018 0
  • Walmart.com / Eng data_007
    “Half of what I spent on ads is a waste... the trouble is I don’t know which half” - by someone, too lazy to google... honestly, I think it may be like 80% wasted, 20% converted.

    Facebook ads are definitely a notch up. They look native unlike pop-up/floating ads on other websites.
    Jul 26, 2018 2
    • Snapchat NRBM24
      OP
      What do you mean by 20% converted? Typically click rates even for pull channels are under 5% and conversion rates are 2-4% of the 5% that click through.
      Jul 26, 2018
    • Walmart.com / Eng data_007
      Ok, converted is wrong. I meant 20% ads may have some subconscious impact on viewers’ mind. For example, when I see Amazon ad, I may not click it but it makes me aware that there is something like Amazon and it is selling items. This is impossible to measure though. It maybe even 10%... I don’t know... but there definitely is impact.
      Jul 26, 2018
  • eBay pWrF21
    Ads are just about brand awareness...
    Jul 25, 2018 2
    • Snapchat NRBM24
      OP
      Tell that to “performance marketing” agencies
      Jul 25, 2018
    • eBay pWrF21
      Doesn’t change the reality
      Jul 26, 2018
  • Uber / Other
    Ehfink

    Uber Other

    PRE
    Uber, Google
    Ehfinkmore
    I did chuckle when I read your post because I know a lot of people at Uber who feel the same as you. And I do understand your frustration at the imperfection of digital advertising attribution / measurement. But unfortunately, that’s just the nature of the business you’re in - this isn’t accounting or pure science, where there is always a black and white perfect answer to everything....marketing will always involve a little bit of art / creativity / psychology/ emotion. Why? Because humans are humans and they don’t always make purchase decisions rationally. And so although I think it’s great to keep experimenting with different attribution models and testing different channels, I’ve seen marketing managers get too obsessed with getting perfect marketing data / analytics....they become Paralysed by their inability to make decisions around optimization’s and channels. So...focus on solutions...not highlighting imperfections and you’ll stay happy in your job, the industry will catch up one day. It’s come along way since the days of pure TV and print advertising .
    Aug 27, 2018 0
  • Google
    vpFb99

    Google

    PRE
    Microsoft
    vpFb99more
    Half of ad spend works. We just don’t know which half
    Aug 7, 2018 1
    • Google
      vpFb99

      Google

      PRE
      Microsoft
      vpFb99more
      😊
      Aug 7, 2018
  • Yahoo Genius
    Honestly your analysis is really good but you missed the bigger point, do people really click on ads and finish transactions (i.e convert). I am mainly speaking about search ads, but same applies to Native ads as well. I never believe there are this many people who really click on ads and buy stuff.
    Jul 25, 2018 3
    • Snapchat NRBM24
      OP
      Surprisingly people do...but in most cases it’s not enough to justify the cost. I.e. last I checked click rates for display on GDN were .021%
      Jul 25, 2018
    • Intel sillycon
      021?
      Jul 25, 2018
    • Snapchat NRBM24
      OP
      .021%
      Jul 25, 2018
  • MZ mht6fh
    Did your agency make you feel like an idiot today or something?
    Jul 25, 2018 1
    • Snapchat NRBM24
      OP
      Haha nope, quite the opposite. I would love that for a change.
      Jul 25, 2018
  • Microsoft Sky net
    Instagram ads are very effective across the funnel primarily because they are high quality compelling images and the format is native
    Jul 25, 2018 0

Salary
Comparison

    Real time salary information from verified employees