Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Uber discovered they’d been defrauded out of 2/3 of their ad spend

twitter.com

Post reply on HN
Print view View original

401–410 of 933 posts

Re: Uber discovered they’d been defrauded out of 2/3 of their ad spend

#401

Post by veeenu »

etempleton wrote: post #292

Because most advertising is not effective on a 1:1 basis. The clean, direct conversion funnel is in most industries and sectors a complete myth. Advertising is about awareness. All good marketing is long-term. Digital advertising has been sold (by overzealous marketers) as magical 1:1 sales machines. It is almost always not that. This is the hardest thing to convey to clients. Clients want a promise and a guarantee o…

I worked in wealth management up until recently and can testify this exact behavioral bias was to be found in almost every single client, and sometimes in professionals as well, especially during market downturns. It's likely to be a mix of several factors, such as the fact that losses have a much stronger emotional impact than gains of similar entity, or the fact that most people aren't really trained to reason probabilistically in an intuitive manner. It is easy to be lured in by someone promising the moon when most people won't settle for anything less.

Re: Uber discovered they’d been defrauded out of 2/3 of their ad spend

#402

Post by simonkafan »

It would be the downfall of the Google empire (and probably half of the Internet) if the majority of companies realized that most of their online/mobile ad budget is absolutely wasted. I don't know of any other industry where people put so much money into it only to have so little tangible evidence at the end of the day of what they really paid for ("look, X real people clicked on your ad.... at least that's what we tell you and whatever that means for you.").

Re: Uber discovered they’d been defrauded out of 2/3 of their ad spend

#404

Post by visarga »

Anon1096 wrote: post #386
Ericson2314 wrote: post #307

Earlier quoted context omitted.

Good. Advertising is a terrible zero-sum industry that just distorts market activity. It has little place in capitalism or any other system. Exceptions: - New products deserve some advertising. I would handle that like the patent system. Demonstrate your product is sufficiently different, and get access to some attention economy time slice. Some may say, well what about competition with existing products to bring dow…

What makes you say advertising is zero-sum? This doesn't seem to be the case at all to me. Take diamonds for example. Prior to the mid 1900s, diamonds weren't nearly as sought after as they are currently. It took the marketing and advertising efforts (among other efforts, like monopolizing the mining process) of De Beers to convince the public that diamonds should be highly coveted. It was advertising that pushed the…

> De Beers is obviously highly derided, but this a pretty clear cut case that advertising is NOT zero sum

If you take into consideration the money wasted on overpriced stones, it's zero sum. How did people ever get engaged before De Beers?

Re: Uber discovered they’d been defrauded out of 2/3 of their ad spend

#405

Post by rocqua »

alacombe wrote: post #302
qeternity wrote: post #194

Earlier quoted context omitted.

No. A negative ROI just implies that the ratio of benefit/cost is less than one. If I incur a cost/investment of $100 but it creates value of $200 then I have an ROI of 100%. If this same expenditure instead only produced $80 value then I would have -20% ROI as in the value I’m realizing from my investment is 20% less than the cost of the investment.

Ratio can't be negative, they can either be above of below 1. What you are talking about is "$benefit - $cost", not "$benefit / cost".

We tend to call most rations below 1 'negative'. Its technically wrong but most people get what you mean when you say it.

Re: Uber discovered they’d been defrauded out of 2/3 of their ad spend

#406

Post by Traster »

mlazos wrote: post #358

Honestly though why tf did these companies never try lowering spend incrementally to see if it made a difference? Is this why engineers are so valued because we understand the simple concept of AB testing?

For the same reason that you never lower your margin. It's very easy to lower your margin, and you might get growth as a result of your lower margins, but it's a massive gamble with the most likely outcome being that you are now selling the same stuff for less money. And do you know what happens when you cut your marigns and don't see growth? You get fired.

Same thing with advertising. You drop your ad spend, you lose market share, and you're fired. There is no step 4. If you're right and the ad spend was unnecessary, you've made a tiny difference (no one cares about cost, only growth), if you're wrong, you're fired.

Re: Uber discovered they’d been defrauded out of 2/3 of their ad spend

#408

Post by texasbigdata »

mlazos wrote: post #358

Honestly though why tf did these companies never try lowering spend incrementally to see if it made a difference? Is this why engineers are so valued because we understand the simple concept of AB testing?

Disclaimer: I’m not a marketing guy.

It’s sold as hand wavy. I once dropped in on a consulting client (at the request of an investor on the board) and did a quick assessment. The assessment was holistic and marketing was one of many enterprise value levers, so did a rough check. The marketing firm, very similar to the Uber commentary in the underlying podcast above where he found 20,000 installs coming from an app with 30,000 MAUs, I calculated claimed performance which back of the envelope would imply you could capture the entire US vertical for that industry for sub $X0M of annual spend. I knew that space very well, and well that just was not possible.

My quick question kicked off a large tap dance in which some finger pointing occurred and a “new attribution formula” was introduced.

The vendor got fired.

This was sort of a reasonably sized entity, with some sophistication. But think about it, unless you’re tech native, your VP of marketing has no granular data abilities in 2020, and unless you have a pool of SQL capable analysts lying around, AND you’re going to insist to validate on your own, it simply doesn’t work.

So you trust the vendor. But wait the vendor gets paid X% of what you spend. Hmmm, would they realllllly work super hard to make sure the results are real if they are not easily found out to be falsified? This is a paranoid stance, obviously. But intrinsically structurally flawed.

The Uber guy on the podcast straight up said he wouldn’t buy a dollar of programmatic ads without a big data team internally and a policy of (loosely quoted) “we need that file to check and if you don’t send it to us we’re cutting you off”.

That’s just a lot of work.

Re: Uber discovered they’d been defrauded out of 2/3 of their ad spend

#409

Post by reilly3000 »

mlazos wrote: post #358

Honestly though why tf did these companies never try lowering spend incrementally to see if it made a difference? Is this why engineers are so valued because we understand the simple concept of AB testing?

Marketers do tons of A/B testing and most measure conversions based on ad copy and placements. The problem often is they don’t look beyond aggregates ( as Google/Facebook flood marketers with dimensions, this one is easy to overlook). So it the overall ROI may be passable, but specific segments are over-analyzed (did YOU place the right message?) and others ignored. The platforms have never ever in any amount of years been willing to hand over raw clickstream logs of ad data, so marketers are forced to rely on misleading aggregates.

I’m sick of engineer elitism. Try harder to see that brilliant folks exist in every department. You’ll be a happier person and learn a lot more. Marketing analytics is fascinating and full of puzzles. Marketers were driving engineers to make multivariate testing scale 20 years ago.

Re: Uber discovered they’d been defrauded out of 2/3 of their ad spend

#410

Post by f430 »

Google is by far the most problematic. You are paying ridiculous amount of money for a click you don't even know came from a legitimate source.

There are many many botnets out there undetected running headless chromium browser instances that is indistinguishable from a real one.

So if you can't tell and then you will never know if you are paying for legitimate prospects or not.

However, advertising for established brands direct response isn't what they are going for rather brand awareness

Post reply on HN