I'm not a honey user but I thought this section was interesting:
> This gets a bit technical but in the video, Jonathon carefully shows you that the ‘NV_MC_LC’ cookie changes from Linus Tech Tips -> Paypal when a user engages with Honey. What he must have seen is that there is also a ‘NV_MC_FC’ cookie that stays affiliated with Linus Tech Tips and is NOT changed to Paypal. In this case LC stands for ‘last click’ and FC for ‘first click’. In the video he seems to claim that there is no first click cookie and only a last click cookie - this claim is false.
> In my DM conversation with Jonathon he claimed that he noticed the FC cookie but didn’t think it was relevant and that he was confused by it. I wonder, as an investigative journalist, did he think to ask anyone at NewEgg or the affiliate networks to explain it to him before he threw damning accusations at an industry he didn’t understand?
I saw that and I'm not convinced this changes anything. The fact that Honey is inserting itself into the affiliate attribution chain at all when it did literally nothing is still wrong to me.
It boils down to making numbers go up. Maybe for you, Honey doesn't do much. But add Honey to the picture, and retailers are seeing an increase in sales and a decrease in cart abandonment. So you choose to partner with a coupon company and pay them commission and for some percentage of users, seeing that popup pushed them over the edge to make the purchase.
In the attribution chain, when you compare an initial referral vs. the coupon app, it's fair to say that the initial referral has more impact. So maybe you want the initial referral to take most or all of the credit. But what about when there's no referral? Doesn't the coupon app deserve to be a part of the chain if it is ultimately driving positive return?
There's a very reasonable argument to be made a number of shopers wouldn't convert if they didn't feel like they were getting a "deal." So honey is undeniably aiding in the sale.
It's not "literally doing nothing" to compile and automatically apply/suggest coupon codes. That's literally doing something. Is it valuable? Objectively, yes, hence the millions upon millions of users.
Your statement is either hyperbolic or disingenuous: the very two things people are accusing honey of doing.
weren't they inserting themselves even in cases where there was no available coupon? Finding the user a deal isn't "doing literally nothing", and the argument for honey inserting itself in that case is at least not crazy. But as I understand it, they inserted themselves in every case even when they very literally did nothing (no deal, nothing).
I do agree that's shady, but there's a good chance at least some of those shoppers wouldn't have converted because they thought to themselves "I'll wait to make sure I can get the best deal" whereas someone with honey was in theory "sure" they were getting the best deal because no deals were available and pulled the trigger. Obviously in some cases that wasn't actually the case, but either way honey DID contribute to the sale.
My point is the issue is not honey. The problem is affiliate marketing (specifically last click attribution) as a whole. Don't hate the player. Hate the game.
The shopper feeling like they've gotten a deal is uncorrelated with whether Honey makes money from the merchants.
They also collect and sell data about all the purchases users make. I'd be startled if PayPal didn't use this data for selling customers on Braintree or selling ads. Also triple dipping and taking money from the affiliates (besides selling user data and extorting merchants) is downright greedy.
That's how any fraud works - the mark feels like he is getting a good deal. And while the fact that there are millions of marks is an objective fact, it's not an objective indicator of the fraud being valuable to the marks.
I fundamentally disagree honey is/was fraudulent. They compiled and distributed coupon codes. The core funtionality exists. It just sucked -- I guess anyway, I was too suspicious and lazy to ever use it myself.
First of all, "undeniably" here is hyperbole. At best you could say "possibly, occasionally". You were already brought to site by a content creator, added the item to your cart, and are in the process of checking out. Why would a coupon code aggregator then deserve the commission for that sale?
> It's not "literally doing nothing" to compile and automatically apply/suggest coupon codes.
Even when they don't find a coupon code, they still take the commission for the sale. That is quite literally the definition of getting paid for doing nothing.
> Is it valuable? Objectively, yes, hence the millions upon millions of users.
Well, no. As the investigation revealed, Honey doesn't actually find any coupon code most of the time. In fact, this is intentional - they partner with retailers to limit the coupon codes they provide to shoppers. In other words they are intentionally providing negative value for the end user most of the time (when compared to searching the Web for a coupon code manually).
You clearly either know nothing about the investigation, or are a Honey employee.
Judging by the condescension oozing from that last statement, it seems you're disinterested in a good faith engagement.
That statement is emotionally charged and factually incorrect, multiple times over. I assume the rest of your reply is along the same lines and won't trouble myself reading it.
I got a weird feeling from the MegaLag video, but overall don't think Honey are entirely in the clear either. From the AMA it seems Honey has been in the business of taking some/all affiliate revenue even in cases where it finds no coupons - sounds like the sites are fine/happy with this, but I'm sure people who post affiliate links are not.
Yeah, the video wasn’t perfect. But honey is clearly a shady business. Honourable businesses don’t need to trick their customers and advertisers about how their business works. Honourable businesses don’t make an enemy of the truth.
Their business model was never explained clearly on their website. Now that how it works has become common knowledge, its absolutely wrecked honey's public perception.
Merchants may have known what they were signing up for (if they signed up at all). But the general public had no idea.
Affiliate attribution wasn't explained in full but MegaLag with all his research still didn't accurately explain it since it's pretty complex. The user doesn't need to know this.
Doesn't make sense to explain all the nuts and bolts if it works the same way any other coupon website would.
This sounds like a distraction. "seems to claim that there is no first click cookie". He brought that up, it doesn't control the payout and doesn't change the result from what I understand. FC cookie is not relevant, Megalag was focusing on what was important information to impart to viewers. If they clicked on an affiliate link from their favorite creator, using Honey hijacked that action of support without disclosing anything.
A potential customer could leave the site, spend time searching for a coupon code, not find anything and give up, and then not return to the site to complete the purchase. Honey could keep users on the site moving through the order flow even if it didn't find anything.
Last Click Attribution is the most common model for affiliate revenue, not First Click Attribution. That the First Click Attribution cookie is still being set is mostly a red herring. Most online sales have 100% of the affiliate revenue going to Last Click Attribution.
He's blatantly ignoring that most affiliate programs only payout to the last-click. Okay...great...the first click attribution is maintained, but if there is no payout for it, then the core issue is still the issue.
Ryan here too - will try to respond to some of these with more info.
The two biggest missing pieces from both my discussion and from the video are:
1) stand down rules for affiliate, and
2) cash back to the user.
I was trying to address the claims he raised in the video specifically and since he didn't mention either I didn't in my reddit post except for a little bit in a couple of the answers.
1)
For the case where the store only uses last click (which is most of them) Honey and other browser extensions follow a rule set by the affiliate networks called 'stand down'. This means they attempt to detect when another affiliate link is clicked (e.g. from a creator) and then either fully disable the functionality or at least don't use affiliate links. Only browser extensions are subject to these rules (e.g. if you visit a coupon code website they will use their affiliate link and override the creator).
Detecting this can be a bit tricky across numerous affiliate networks and I suspect the NewEgg example was selected because it used a non-standard way to manage affiliate tagging and therefore wasn't detected by Honey's stand down logic.
fwiw I agree with the sentiment that Honey shouldn't have been tagging on a 'hey we didn't find any codes' or 'use paypal' click and I personally wouldn't have approved that, though it probably technically does meet most of the affiliate network stand down rules (well, at least it did - I'm sure they've been updated which is a good thing).
2)
Jonathon's video is completely silent on the other core value proposition of Honey: cash back. Honey, like Rakuten, Capital One Shopping, etc, offers cash back funded by affiliate marketing. The model is not new - Ebates (now Rakuten Rewards) was founded in 1998. Honey added this program in 2015.
When a user is shopping with Honey on a store with affiliate commission, Honey almost always gives the user cash back. There are a limited number of exceptions, generally because of the store's policy, and occasionally because there are so many exclusions to the affiliate program that it makes offering cash back confusing to a user.
A valid question to ask is: if a user clicks a creator affiliate link AND has a cash back tool like Honey or Rakuten should they or should they not be eligible for cash back. Personally I think absolutely yes, the user's preference is the most important. But I've heard reasonable people argue the opposite.
What I don't think is that offering it's users cash back makes Honey a scam and I think Jonathon was negligent in presenting this narrative without even considering this primary use case for what is actually the #1 business model in affiliate marketing.
I'll stop there. Happy to answer a few more questions here.
Isn't the problem that none of this was transparent to the user? That honey takes affiliate commissions and MAYBE gives some of that money back to the user? That honey takes credit for the sale and MAYBE stands down in certain circumstances?
I don't think any of this is transparent to the user. That's the scam.
We always tried to be as transparent to our users as possible in the product, in the faq, in our customer support, etc.
You can see evidence of that approach at 6:17 in Jonathon's video in the response that he got from customer service about how Honey works (even when he intentionally removes critical context). He reads a support email that says:
"If Honey is activated and is the last program used while shopping on a site, it is likely Honey will receive credit for the purchase, and Gold will be earned by the member. However, if your favorite influencer's affiliate link was the last program associated with your purchase during your shopping on the site then they will receive the credit for the purchase. Keep in..."
Notably he stopped reading before the "and Gold will be earned by the member" because it didn't fit his narrative.
I know it flashed on the screen faster than anyone could read and was zoomed so you couldn't see the whole thing on one screen but would you consider this level of transparency adequate if he read the whole thing?
> A valid question to ask is: if a user clicks a creator affiliate link AND has a cash back tool like Honey or Rakuten should they or should they not be eligible for cash back. Personally I think absolutely yes, the user's preference is the most important. But I've heard reasonable people argue the opposite.
I'm not very familiar with Honey, it's business model, or the ins and outs of the affiliate program back stage. My impression here is strictly as an observer with no foot in the race.
Should they be eligible for cash back? Sure, if you want to give them cash. I don't see why that entitles Honey a link in the affiliate program chain, though. The cash back offer seems entirely independent from the user's choice to at least get to the point of purchase, with maybe an occassional situation where they would have pulled out if not for a cash back or coupon deal.
It's a benefit you're affording the user, but it's on you to find a way to monetize that without impacting the actual affiliate. If you have a deal with the merchant so that doesn't affect the actual affiliate, great. If you charge your users a monthly fee or something, wonderful. But if you're deciding that you deserve a share of an existing pot, you're in the wrong, and there's no two ways about that. A flawed business model doesn't entitle you to other people's shares.
Again, I'm not familiar with the way the system works at all, and haven't seen Laing's video, so I might be missing context. But from your quote on Reddit:
> On most stores Honey (and others) offer a portion of the commission back to users as cash back.
My understanding is that Honey inserts itself into the affiliate chain, takes Y% of the commission (and reducing the original affiliate's commission by the same amount? Clarification needed), then returns X% back to the user, and keeps Y% minus X% for itself. So what exactly is Honey doing here? Taking a part of the pot because it's giving some portion of the pot back to the user, while otherwise offering nothing of value to either the affiliate or the merchant? Why shouldn't the original affiliate simply be given the ability to offer the user cash back and remove Honey as the middleman? Why should the original affiliate have any loss in their own commission because of a 3rd party's actions?
If Honey's %Y commission is part of a deal you have with the retailer and doesn't affect the original affiliate's commission at all, I apologize, and understand the situation. But if there's any cost to the original affiliate here whatsoever, I don't think you have any justification for imposing that cost on them.
Probably too complex to explain here but tldr; every cash back program is built on top of the affiliate marketing rails since Ebates started it in 1998. Same for coupon websites since those came along.
It is a system design flaw (that maybe will self correct because of this) that multiple advertising models at different points in the value chain are built on the same system. 'Multi-touch' or 'any-click' are the correct direction to solve this problem but introduce their own challenges for retailers which is why most of them have not adopted these systems yet.
You ask "what exactly is Honey doing here?"
Short answer is helping the retailer with conversion and limiting cart abandonment by making the user happy and more likely to transact.
https://old.reddit.com/r/IAmA/comments/1jlfms8/im_ryan_hudso...
I'm not a honey user but I thought this section was interesting:
> This gets a bit technical but in the video, Jonathon carefully shows you that the ‘NV_MC_LC’ cookie changes from Linus Tech Tips -> Paypal when a user engages with Honey. What he must have seen is that there is also a ‘NV_MC_FC’ cookie that stays affiliated with Linus Tech Tips and is NOT changed to Paypal. In this case LC stands for ‘last click’ and FC for ‘first click’. In the video he seems to claim that there is no first click cookie and only a last click cookie - this claim is false.
> In my DM conversation with Jonathon he claimed that he noticed the FC cookie but didn’t think it was relevant and that he was confused by it. I wonder, as an investigative journalist, did he think to ask anyone at NewEgg or the affiliate networks to explain it to him before he threw damning accusations at an industry he didn’t understand?