by nsmog767 on 10/13/21, 4:03 PM with 347 comments
by simonsarris on 10/13/21, 5:19 PM
> [picture of that search term and williams sonoma ads with shopping links]
The main problem here is that if Williams Sonoma was not advertising on that search term, Lodge and Food52 etc etc would, and then those companies would be above the Williams Sonoma organic placements.
The spend is necessary in a defensive way because Google creates a bidding war even for the hyper relevant.
edit: I just checked and if you search "williams sonoma skillet", if WS was not paying for [green] then the very first "result" (ad) would be Food52 [red] https://imgur.com/a/9Nnxs6h
I just tried "airbnb paris" and the first result is, somewhat predictably, an ad that is not airbnb. But the second one is also an ad, this time from airbnb. So they clearly didn't keep their spend dialed down to zero, and are aware of the need to advertise on their own keyword.
by aerosmile on 10/13/21, 6:52 PM
But after reading this article, it finally dawned on me. He makes imperfect conclusions in everything he touches, it's just that in some fields those conclusions can be more easily proved to be wrong than in others. SEO is the perfect field where a polished presenter can get away with imperfect conclusions for years - trust me, I know, I made a living for years in this field, and I am very familiar with the nature of this work. Most of the time, you have no idea what the black box really does, and instead you're just trying to guess what might have happened. Most importantly, there are many ways to skin a cat in SEO, and just because your approach is net positive doesn't mean that you truly are delivering the global maximum (or that the net positive gain was ROI positive). In short, it's impossible to know who's right and who's wrong, and Rand's videos convinced me that he's right, but I am no longer sure. I just rewatched one of them, and can easily see how his conclusions are just... opinions.
While we may or may never find out if his SEO opinions were the global maximum, we can quantifiably demonstrate that his opinions on content marketing are not solid. This whole essay he wrote can be replaced with "hey performance marketers, don't trust the platform numbers and instead do your incrementality studies." Platforms like Facebook will give you those for free if you reach a certain spend level, and you can also get them from 3rd party providers like measured.com. In other words, if you're a performance marketer and you're not conducting incrementality studies, then you're very early in your career and are not following the best practices. Simple as that - no need to extrapolate from there and reach all sorts of additional conclusions (which is obviously a pattern in Rand's behavior) - calling into question a perfectly investable marketing channel, conflating the needs of a public company with everyone else's needs, using words like scam, etc.
I am really disappointed to have to write this, but you would have been better off not reading this article. If Rand is really advocating that the majority of entrepreneurs should follow his advice and focus on PR instead of performance marketing, then perhaps an honest thing to ask would be - how is that working out for his own company? AFAIK, SparkToro is nowhere close to replicating the growth of his previous company, which is honestly disappointing for someone with such a huge reach and name recognition.
by ssharp on 10/13/21, 7:14 PM
#1 - Eliminated all paid search other than some limited branded search terms and shifted all the money to affiliates who were way better at making profits on the keywords we were competing on
#2 - Eliminated all display advertising after running numerous experiments showing it provided almost not incremental conversions, even though the platforms happily took credit for them.
Those two things drove our blended CAC down substantially and by building better affiliate relationships, sales actually increased.
The lesson here is that you need to try a lot of things out and you should be continuously questioning what you're doing and running specific experiments to gut check effectiveness of any ad platform that is slapping cookies on wide groups of users and claiming conversions.
My suspicion is that this is near impossible at any large organization, even one as new as Airbnb. I can just imagine someone walking into a team of 20+ performance marketers and suggesting they need to experiment to determine if any of it is remotely effective. COVID forced them into this but it's something that they should have already been doing.
I also suspect that the top line focus/obsession of most VC-backed companies make this type of exercise seem almost counterintuitive. Don't mess with or question the momentum.
by rexreed on 10/13/21, 9:17 PM
We're already seeing that. The quality of Google products continue to decay. Facebook and LinkedIn are increasingly both becoming shallower advertising hustles (LinkedIn just this week turned off post notification for events to force people to buy LinkedIn ads). As other apps and websites get snapped up by these FAANGs, we'll start to pine for the Internet that was 2008. The decay is already well under way.
What is truly sad is that some of the smartest computing minds of our day are spending their efforts at these FAANGs not advancing society but rather helping keep people more addicted to social networks, optimize for clicks on video and web streams, pushing products in all your channels, and optimizing for the wrong things. How have we gone so astray?
by netcan on 10/13/21, 8:09 PM
Meanwhile, it's a marketing cliche that "half the budget is wasted, but we don't know which." It's also true that google or FB provided analytics, using default settings often grossly overestimate ad effectiveness. All true. A journalist somewhere is writing a version of this article at any given time.
But... From the merchant's perspective, the existence-proofs for advertising's effectiveness are undeniable. Launch a site. No visitors. Advertise. Now there are visitors. People subscribed to something or bought something. The ROI may or may not work, but the principle isn't in question.
For a blank slate, newly launched business performance marketing is easy to measure precisely and you can have a reliable ROI. For BMW, GoPro or geico insurance... the world is more complex, ROIs are more theoretical and "half the budget is wasted, give or take 50%" applies.
The same was true for TV. A mattress store run ads with a crazy guy screaming "Sale!" and the next day a lot of mattresses get sold. The fact that ads made people come buy mattresses is trivially true, from the merchant's POV.
by tomhoward on 10/14/21, 4:35 AM
It didn’t make sense to me that it could be so attractive to brands to promote coupons there. People have already decided they want to buy the product; why would you want them to go hunting for a big discount after they’re already sold?
He told me that the people who like it are the advertising staff and consultants, so they can generate evidence for their bosses/clients that their campaigns are working.
I get that it can be partially effective. Sometimes I’ll think “I’ll buy this product if I can find a coupon that gets the price under X”.
It’s just a funny old world when an advertising professional is motivated to spend money with a third party to give their customer a big discount on a product they’ve already decided they want, in order for the advertising professional to justify their existence.
by s17n on 10/13/21, 6:00 PM
by Ozzie_osman on 10/13/21, 5:42 PM
One additional thing people don't call out is that a lot of the budgets spent on these platforms are "learning" budgets. Agencies play this card really well. They'll tell you, "oh, you need to increase your budget, and test all these different combinations of ads/targeting/landing pages/etc so that you can learn what works (or the AI behind the platform can learn what works)". And obviously, in "learning" mode, you're ignoring the ROI.
I've seen people spend substantial amounts of money in "learning mode", and the platforms are kind of incentivized to make the learning less efficient so it takes longer and more spend for you to get to ROI positive (or to learn that you will never get there).
by myth2018 on 10/13/21, 10:30 PM
My ads were working reasonabilly well considering the low investment I was making, with a fair amount of prospects filling my contact form on a low but steady rate. I was satisfied with the return I was getting.
However, the pandemic caused a significant drop on my product's demand. I thought I was going to get little to no contacts from the moment the lockdown was announced on, but:
1) I kept getting clicks at basically the same rate -- therefore my budget kept being depleted as it used to be;
2) Bounce rates increased A LOT;
3) The few actual people who got in touch were not actually looking for the product I announced, but similar ones (which I didn't announce nor sell);
So, according to my experience, I can't say adwords totally doesn't work.. but I'd say their algorithms are optimized to spend your money regardless of the results you're going to obtain.
by citizenpaul on 10/13/21, 7:46 PM
I once even had a marketing department shut down with analytical proof. I got tired of the marketing department with probably 50x the IT department budget constantly jumping down our throats about how "IF ONLY IT HAD DELIVERED X BY X we could have had 100,000,000,000x /s sales this month"
I made a dynamic report dashboard in my first react project to analyze market spending and prove that even if you wanted to move around metrics to be comically generous the marketing was doing basically nothing to drive sales. MGMT got rid of them and literally nothing changed except everyone had better budgets.
by kposehn on 10/13/21, 6:08 PM
> Reliably someone comes along every few months to question [performance marketing]. I always come back to analyses of incrementality as the real proof.
> Take an audience of X people. Divide them in two. Show ads to your test group, don't show to control. Watch your business grow and gauge the lift between the two audiences. The companies that know how to advertise at scale do this constantly and can gauge the real effect of their ad dollars. Facebook, Google and others make these tests possible in their platforms, while other software suites such as Impact Altitude and VisualIQ allow you to do this kind of analysis and testing as well.
> In the end, most of it proves out to be incremental. There are notable exceptions of course, but when are there not?
by cwp on 10/14/21, 12:21 AM
This is really the sweet spot for digital advertising. The cost is low, both because you don't need that many new users to make a huge difference, and because the volume is so low that the targeting works really, really well, and conversion rates are really high. For a more established business, sure, you're getting people that would have bought anyway. But in this case, you're getting people that would have bought if they knew you existed, which they don't.
by missedthecue on 10/13/21, 6:13 PM
"Rental stays in XYZ city" will bring up an airbnb result all day, probably in the top five results. Therefore, paying $3 a click to be placed above your own search result is probably silly.
I happen to have a small side project and advertise it with a very low budget on facebook, google, and bing. It works. I don't rank very high because my SEO skills are poor, but google ads absolutely drive real and interested people to my site.
by arbuge on 10/13/21, 5:36 PM
(If you have a near-monopoly on planners, of course, such an ad would indeed have been a waste. The author would have come to you by default once he decided to get a planner. )
by mkmk on 10/13/21, 5:31 PM
Hard not to see this as a clumsy sales pitch for his company, especially when it starts with such a disingenuous example/quote (of course AirBnb didn't have to spend on performance marketing during a global pandemic where everybody was suddenly looking for a getaway...)
by redelbee on 10/13/21, 6:43 PM
So if you’re buying or even just evaluating performance ads without considering the bigger picture you might come to erroneous conclusions.
Take the Lego Movie example from the article. The $65 million movie is no doubt an awareness play. Could you make the case that you should also increase your performance budget to help capture more of the demand you just generated with the movie? Or should you just hope that people go from the movie theater to buy Lego unprompted? Is it worth it for Lego to advertise to people who walk out of the theater and search for “Lego Batman set” or whatever? I think so, even though evaluating such branded search campaigns individually might make them seem inefficient.
It seems very easy to dismiss the performance advertising as a scam when you evaluate it in a vacuum. As noted in the article it’s important (and very difficult) to understand the incremental outcome of any channel or campaign. That incrementality includes awareness campaigns.
After more than a decade in advertising and marketing I am now more than ever unwilling to accept simple or definitive answers to highly complicated questions. At best I hope that we can unwind some of the overall complexity so we can have a chance to trust some of those definitive answers.
by projektfu on 10/13/21, 10:09 PM
by avalys on 10/13/21, 5:29 PM
by andrewyates2020 on 10/13/21, 8:51 PM
However, every advertising platform will take your money if you give it to them, so buyer beware.
True incrementally is a challenging measurement issue and therefore even more challenging to predict for delivery via ML. It's real, though, just hard.
Buyer beware. When buying ads, think the stock market. If you hear stories about 20x ROI from random people, do what you when you hear 20x ROI on stock picks: nothing.
by toinbis on 10/13/21, 8:47 PM
Yes, you do have a challanging problem of attribution. But the spend and revenue figures are what matters at the end of the day. And neither of them has any area for scamming (let's ignore edge cases).
Disclosure: only skimmed through the article and my arguments above are just directed towards the headline. However credible and opinion leader the author - Rand Fishkin - is, the article itself at the first glance did not inspired me personally as a worthy my attentive reading time.
by stephc_int13 on 10/13/21, 7:11 PM
I think this is overall a waste of resources and I'd like to see a more virtuous system, but I fail to imagine one.
by carbine on 10/14/21, 2:58 PM
I’ve been in growth for years and while I dislike PPC and find it boring and obfuscatory (FB and Google will both happily claim credit for the same purchase), there’s no denying that in many cases it works. I’ve been at early stage ecomm co’s where we’d pause the spend to see what happened and — what do ya know? — sales would plummet. Brands relying on it for discovery often benefit substantially.
I’d argue that most of the benefits of PPC accrue to smaller businesses (and the platforms, obv). The error in our collective framing is to hate on advertising because we think of it as coming from big brands. But these days it enables vastly more entrepreneurship and competition than I think many on here realize.
And to everyone cheering on Apple’s recent changes that make FB ad tracking impossible — who do you think suffers as a result? Big brands like LEGO can just redirect funds to a Hollywood film or some other high-visibility activation, because they already have their customers’ attention. It’s the small brands paying the price.
by CaveTech on 10/13/21, 6:13 PM
The whole point of _performance_ advertising is that it's effect is _measurable_. If AirBnB spent $500m+ on performance advertising, they should be able to trace that back to an exact amount of revenue. If you are a brand in this scenario, you can conduct split tests by sampling the conversion rate of users from advertising vs non-advertising. Again, it should be simple to see if the conversion rate for users targeted via advertising has increased or is unchanged.
In the branded search examples, again, as an advertiser, you can see what searches are associated with your leads. While you do have to compete for attention on your own branded search (to compete against competitors taking the slot), you should also be able to recognize unbranded terms which drive conversions for you. Again, assuming this is actually _performance_ marketing, you would be able to look at the cost of these placements and the ROI, and the impact would be measurable.
The rest of the article is largely composed of straw-man arguments that imply the results are not measurable, when in fact they are (if done right).
disclaimer: I'm the CTO of a performance marketing platform. The vast majority of conversions on our platform happen same-session. There's a very easy way to measure this effect -> pause your campaigns and immediately see conversions fall.
by rapht on 10/13/21, 7:43 PM
On the other hand, while I do indeed believe that the "ROI" from Performance Advertising is something between just false and deliberately misleading, the bigger picture that I'm interested in is marketshare. Because when looking at market share, it's not a question of incrementality anymore, but whether you're growing slower/faster than your competitors, and your cost of doing that, and at what point you're OK to 'buy' marketshare, in the sense of losing money in the pursuit of growth, and how much. And then, OK, let's talk about ROI on that basis - most of the time, achieving this will indeed require tools from the Performance Adversiting toolbox, which allow you to conveniently track the amount of marketshare (i.e. sales) you bought.
by rfwhyte on 10/13/21, 8:50 PM
Any article that says "Don't buy X buy Y" loses a lot of credibility when it's written buy a guy who sells Y.
by happybuy on 10/13/21, 11:32 PM
When you start using these tools, they almost seem to replace what would have been an organic acquisition with a paid acquisition.
As the article notes, why isn't this obvious for most large advertisers on these platforms?
Likely due to:
- Their spend is so large that it hides the fact that they would have gotten a large % of it for free anyway
- Large advertisers rarely stop all advertising for a period of time to notice the % of truely organic traffic (that isn't just opportunistically attributed to the platforms)
- They have been doing it for so long that big spends on these platforms are considered mandatory and not even questioned
by adotjdotr on 10/14/21, 9:25 AM
So the idea it is a scam is stupid.
by tialaramex on 10/13/21, 11:32 PM
Some years ago, I treated myself to some nice Sophia Webster shoes.
I'm a tall man, my feet don't fit in any of her shoes, but a colleague had pointed out that er, they're for looking at, and if they're on my feet it's actually harder for me to look at them, so, why not buy them and put them on a display shelf like I would a nice sculpture. That made good sense. I bought an arbitrary size that was available - it's way easier to get the shoes you want if they don't need to fit.
Now of course as a tall man, even though it's not unknown for me to buy heels (in a size I can actually wear) it's rare, and I don't buy skirts, have no use for a bra and so on. But still, advertising isn't very smart, so it's not a huge surprise that after buying those shoes I got considerably more adverts for stereotypically female clothing and accessories even on platforms that explicitly know my gender and previous purchasing habits.
What did surprise me is how many adverts I got for the exact pair of shoes I had just bought. Not similar shoes from other designers. Not other shoes Sophia Webster has designed - the exact identical pair of shoes, often from the exact vendor who just sold them to me.
And my theory as to why is pretty simple. Advertising doesn't understand time's arrow. If you only ask whether these adverts are associated with my buying the shoes that's true. They showed adverts to me, I bought the shoes. The only problem is that causality is reversed, I bought the shoes and then they showed me the adverts.
This was confirmed to me by adverts the grocery delivery company I sometimes use ran at that time. The adverts specified the exact goods I buy from them, perishable goods which I had already booked a delivery of, saying hey, why not buy these from us. Well, I already did, duh? But so long as nobody remembers to analyse whether my actual purchase decision (not the delivery) was before or after the advert these adverts probably look like they have amazingly high conversion rates and you couldn't have targeted a traditional advert so narrowly, it must be worth it...
by morelandjs on 10/13/21, 10:12 PM
by tomlockwood on 10/14/21, 1:14 AM
https://lockwood.dev/advertising/2019/06/07/adtech-sucks.htm...
https://lockwood.dev/advertising/2019/06/14/adtech-sucks-thi...
by webmobdev on 10/14/21, 3:56 PM
by ChuckMcM on 10/13/21, 7:02 PM
[1] or black hat SEO for that matter.
by ckelly on 10/14/21, 6:52 AM
by fungiblecog on 10/13/21, 9:05 PM
I'm constantly bombarded by ads for stuff I've already decided to buy (or more usually that I already bought last week). This stuff has zero value. The point of advertising has always (until now) been about taking your product to people who didn't know it was there. Targeted advertising would be a funny joke if so many people didn't take it seriously.
by timdellinger on 10/13/21, 6:51 PM
by idontwantthis on 10/14/21, 3:44 AM
by fourseventy on 10/13/21, 5:32 PM
The most common source of inflation in Google/FB self reported performance numbers is multiple ad channels taking credit for the same order. If a customer clicks a Google ad then clicks a Facebook Ad then makes a purchase, each ad channel will claim credit for that purchase. In reality each ad platform only has claim to ~50% of the purchase (depending on what attribution algorithm you want to use).
In terms of knowing that real value is produced from these ad channels, I see it every day in many of our customers data. Clients will increase or decrease ad spend and there will be a correlated increase/decrease in sales.
tldr; Google/Facebook ads over report their numbers but do ultimately drive sales according to the data I have first hand access to.
by DevKoala on 10/13/21, 7:11 PM
by richardfey on 10/14/21, 6:27 AM
by kornhole on 10/14/21, 1:17 AM
by cblconfederate on 10/13/21, 5:43 PM
by xiaodai on 10/14/21, 9:32 AM
by xiaodai on 10/14/21, 8:39 AM
by xlance on 10/13/21, 9:57 PM
Performance marketing = Distribution. Without performance marketing your competitors take the sale instead.
by ulises314 on 10/13/21, 6:37 PM
by chmsky00 on 10/13/21, 6:06 PM
It sounds “deep state” but it’s actually plainly documented in government files and written about by reliable sources.
Remember we’re still emerging from an era of whispering the same old story of morality and obligation to each other.
I am not at all interested in helping someone build a fertilizer empire or pillow brand. Politically my hands are tied to doing so if I want a life.
Poor people effectively live a life of quota and state sanctioned limits on their access to material support by cutting social programs with public support.
Advertising America as anything but a sanctimonious police state is a scam.
by rvnx on 10/13/21, 5:34 PM