Removing email registration improved retention
solitaired.com
solitaired.com
The most successful and lucrative form of marketing, by far, is re-marketing.
That is where you take someone who signed up but is not currently a customer and you target Facebook/Google ads specifically at that email address. I've seen conversion rates as high as 30% and it's typically pretty affordable.
It's such a critical part of marketing that many companies will take a loss on the initial "here is our product, please join" ads just so they can follow up with re-marketing in the weeks/months to come. And because people re-use their email addresses across websites you can target them on Quora, Reddit etc as well.
You can't do any of this without their email.
That is scummy as hell and might even get you in trouble when it comes to the GDPR if you're operating in the EU.
If I sign up for your web service the last thing I want is Facebook/Google knowing that fact.
You've signed up for a web service and never seen ads on other sites for it ? Very strange.
So yes, technically you can ask the user for consent, but it has to be explicit ("we'd like to share your e-mail/phone number with our advertising partners such as Facebook, accept/decline?") and I can't imagine anyone in their right mind consenting to that.
> You've signed up for a web service and never seen ads on other sites for it ? Very strange.
I sign up for stuff only when I have no other choice for exactly this reason, and often provide fake details. Reminds me of an ex-client where they had an issue with their potential customers not providing the right contact details because they're afraid we're going to spam them. "But do we actually spam them? -Yes."
If I'm not on Facebook (which I'm not) you are telling them that there most likely exist a user with this email address and an interest in your service. If many companies do this FB might even be able to build a profile of me without me doing anything
This is (or at least should be) not Bueno under GDPR / data minimalization.
If Facebook is only using something like a hash of an email address in order to target ads at specific Facebook users at the request of one of their advertisers, they are probably only acting as a data processor for a very specific purpose that might be acceptable for both Facebook themselves and the advertiser under the GDPR rules.
If Facebook does anything else at all with that data, their role probably changes from a GDPR perspective. The hash is personal data, since by definition it's being used to identify a specific person. If Facebook is using the data they have associated with that hash -- for example, anything they know about the business that provided it -- to build up more of a profile on their users, they are probably now a data controller, possibly as well as a data processor in connection with the original targeted ad process. Then you get into questions about whether Facebook's users have given their suitably informed consent to Facebook or there is some other lawful basis for whatever processing is happening.
Obviously if businesses were providing actual email addresses to Facebook or if Facebook were using that data to do things like building shadow profiles on non-Facebook users, that would be another level entirely. And AFAIK, the custom audience tools on marketing platforms like Facebook typically do accept directly uploads of literal email addresses, phone numbers or other identifying details for the audience to be targeted, so maybe the discussion about hashing above is all moot anyway.
I think Facebook hashes any plain text you give it client side to (try) to buffer their defenses... but I’m not sure. It’s a theater.
I wrote a blog post about this data sharing with Facebook https://blog.gingerlime.com/2020/whos-sharing-my-data-and-wh... that digs a little in this direction.
You're also telling Facebook "by the way, I have a relationship with someone with this email address". That's personally identifiable information that you're sending to Facebook. Under the GDPR you can only do that if you have the explicit and freely given opt-in permission to do that from each respective person. "By using this site you agree to..." or "by signing up you agree to..." does not qualify as consent under the GDPR.
If the person does not live in Europe and you are not in Europe then the GDPR doesn't apply, of course.
https://www.lda.bayern.de/media/pm2017_07.pdf is the official press release of the Bavarian DPA (in German), with their guidelines attached (starting on page 4).
There is two kinds of "custom audiences" - one list-based and one based on tracking pixels. I'll only quote the parts relevant to the method where customer lists are uploaded.
a.Rechtmäßiger Einsatz - Der Einsatz ist nur aufgrund einer informierten Einwilligung der Kunden zulässig. Das Hochladen der Kundenliste kann weder auf eine Rechtsgrundlage des BDSG noch des TMG gestützt werden. Diese Rechtsauffassung beruht auf einer europarechtskonformen Auslegung der geltenden deutschen Datenschutzbestimmungen und berücksichtigt die jüngsten Entscheidungen des EuGH zum Datenschutz. Im Übrigen wird das Übermitteln dieser Liste an Facebook auch auf der Basis des ab Mai 2018 geltenden Rechts, d.h. nach der Datenschutz-Grundverordnung (DS-GVO), nicht ohne Einwilligung zulässig sein.
b.Widerruf der Einwilligung - Widerruft der Betroffene seine Einwilligung, so muss er von der Kundenliste entfernt werden. Da der Webseiten-Betreiber keine Kenntnis davon hat, welche Kunden auch Nutzer auf Facebook sind und beworben werden, ist die vollständige Custom Audience-Liste unverzüglich zu aktualisieren.
(Translation - Google translate with misleading issues corrected manually:
Lawful use - Use is only permitted with the informed consent of the customer. The uploading of the customer list can neither be based on a legal basis of the BDSG nor the TMG. This legal opinion is based on an interpretation of the applicable German data protection regulations in accordance with European law and takes into account the most recent decisions of the ECJ on data protection. Beyond that, transmitting this list to Facebook will also not be permitted without consent according to the law applicable from May 2018, i.e. according to the General Data Protection Regulation (GDPR).
Withdrawal of consent - If the person concerned withdraws his or her consent, he or she must be removed from the customer list. Since the website operator has no knowledge of which customers are also users on Facebook and are being advertised, the complete Custom Audience list must be updated immediately.)
Questionable. I guarantee the vast majority of users don't even read the massive legalese text walls companies show them before they sign up. Usability studies have shown that people don't even read small error messages, they just want to get rid of the annoying message as quickly as possible. The few of them that actually do read these things probably won't have the foggiest idea what any of it means or the risks associated with the breach of their privacy. So how could this be real informed consent?
Of course, we also have sites where this document is not shown at any time and can only be reached through a link buried in the page's footer. Sites that just write whatever terms they want into this hidden page and then say everyone is agreeing with it by virtue of using the site.
Under the GDPR, any non-essential data processing (analytics, ads, marketing, etc falls into that) should be opt-in and dark patterns like pre-ticked checkboxes are not allowed.
This isn't strictly true. Consent is only one lawful basis for processing under GDPR, and it comes with a lot of strings attached that other bases don't necessarily have, which is why so many lawyers and consultants were recommending against relying it unless it was the only way during the mad rush to GDPR compliance a few years back.
In particular, even some of the regulators have themselves indicated that marketing might be a legitimate interest of a business. Obviously the details matter here, and handing personal data over to third parties like Facebook without their knowledge or consent seems materially different to, for example, the original business sending a relevant email about a new product that is related to something that the recipient already bought from them. Time will tell how the regulators decide to handle this.
I agree with you on this part. It is not a violation of GDPR on the ad platform side since you, as the data controller, are responsible to obtain a permission from the end-user. The ad platform is a data processor defined under GDPR. I am sure that the agreement between you and the ad platform is stating that you have a permission to use the email addresses for targeted advertising purposes and bear the full legal responsibility if not.
> since users are giving consent when they signup.
See Nextgrids comment. Yes, the GDPR admittedly lacks on the enforcement side and yes, I agree that this is a common practice, but that does not make it legal. Not for a data subject residing in the EU.
Is it? I've never seen such ads. Or any ads for that matter, since every device and browser has adblock these days.
I think we'll see regulators take a different view when they get around to challenging this practice, and the businesses who get made into examples might find it an expensive lesson. Handing over personal details to big data hoarders for remarketing purposes is the epitome of behaviour the GDPR was intended to curtail. You can't just mutter the word "consent" and claim some small print on a Ts & Cs page no-one reads protects you, and regulators have shown very little sympathy so far for data controllers who have tried to weasel their way out of GDPR obligations with this kind of strategy.
Those regulators are still under-resourced and it will presumably take some time for them to get around to dealing with this issue. Right now they're still going after serious leaks and the like. But they're already handing out 9-figure fines to big name businesses for those breaches, and by default those fines go back into central government coffers. Given the current economic climate, how long do you think it will be before their governments realise that this is potentially a very lucrative revenue stream that the public is unlikely to mind, and so start pushing the funding for those regulators up? The ICO (the UK's regulator) has already significantly increased its budget and headcount since the GDPR came into effect, and is reportedly looking at ways to ringfence some of the fines to cover the litigation costs when it inevitably has to defend the big penalties it will hand down from time to time.
When the Cambridge Analytica scandal happened here in the UK, the ICO fined Facebook £500,000. That was the largest fine they could legally impose at the time. As they observed themselves, in what might charitably be considered a thinly veiled threat, under the GDPR that could have been well over £1B instead. Even an organisation the size of Facebook is going to feel that, particularly since there is nothing that says it can't be repeatedly fined on that scale if it misbehaves in multiple different ways.
A couple of potentially important issues have, as far as I know, not yet been resolved in this area.
Firstly, what happens if processing in violation of the GDPR is widespread, the businesses you give your address to are the data controllers, but you still have the likes of Facebook hoovering up huge amounts of personal data inappropriately but possibly only in a capacity of data processor? No doubt there will be some interesting legal arguments about where liability is going to be placed if Facebook was actively soliciting that sort of activity as part of its business model.
Secondly, what happens after the UK has fully separated from the EU at the end of this year, if as the government has stated we retain the GDPR in our national law? Until Brexit was relevant, the GDPR was an EU-wide measure, and typically one member state's regulator would take the lead role in any given case. Anyone breaking the GDPR's rules could be duly investigated and penalised, but only once, not in the same way by every regulator in every member state where there was offending behaviour. If the UK is no longer to be a part of that scheme, will regulators still co-ordinate in this way, or will the businesses sharing data with Facebook face a kind of double jeopardy where both the UK and a lead regulator from an EU member state can potentially fine them for the same behaviour, effectively doubling the maximum penalty they could receive?
If both of those issues were resolved in ways unfavourable to the marketing platforms like Facebook, they could be looking at huge fines for promoting this sort of scheme on the scale that they do, potentially enough to make whole strategies based on selective targeting unviable.
That's not how it works. Hiding the "consent" in the fine print doesn't count, and at least in Germany, it's clear that you need valid consent and can't weasel out of it by claiming "legitimate interest" etc.
I already had a DPA explain this to one of the companies that decided to give my data to Facebook, and the DPA indicated that they were acting on multiple complaints in that regard.
There's a good chance they'll let you get away with a warning the first time if you haven't gotten in trouble before, but especially if you keep doing it (or if they decide that by now, you certainly should have known), expect quickly escalating fines.
If advertising was targeted at the browser level (the browser has access to the entire catalog of ads out there and then does the selection locally based on sites/services I interacted with previously) then I would be in favor of that.
Finally you are omitting a third option in your comparison: how about no advertising at all? Preferring paid services over ad-supported ones and countermeasures like uBlock Origin make that a real possibility. I can't recall the last time I've seen a proper ad online (in fact my problem with the parent's idea is more about the data sharing than the ads themselves since I won't see the ads anyway).
No advertising isn't a viable option in this world. I'd go as far as to say that the Internet, as we know it today, would not exist without targeted ads.
Society has laws for a reason when its constituents decide that certain behavior is detrimental to it and should be outlawed & discouraged by the use of appropriate punishment. I don't see why this shouldn't apply here? The GDPR is in fact a step in that direction, though its enforcement is severely lacking.
> No advertising isn't a viable option in this world.
This is debatable but it's a discussion for another thread.
> I'd go as far as to say that the Internet, as we know it today, would not exist without targeted ads.
The Internet originally was about sharing information freely. It facilitated commerce to a certain extent but commerce wasn't its core purpose. The internet as well have nowadays has actually become worse because of the increased focus on commerce & advertising.
Criminal activities hurt people. Ads do not.
Especially in this case where PII data is not being provided to the advertising company.
They do. Advertising's whole purpose is to manipulate people into doing what companies want them to do. Not every company has your best interests in mind. Due to this inherent conflict of interest, ads should be viewed with healthy suspicion at best.
Banning tobacco ads helped reduce smoking in my country. We should ban a whole lot more.
There are plenty of scams and malware being spread through ads. Furthermore ads are a parasite that wastes most people's time for no benefit with no official way for them to opt-out (a lot of services don't allow you to pay money to opt-out); it'a a cancer on society.
> Especially in this case where PII data is not being provided to the advertising company.
You are literally talking about capturing e-mail addresses so you can pass them to an advertising partner to target ads to these users. How is that not PII?
As yourself this: How do you feel about the possibility of any personal information you give to any company may be given to others without your consent (or "with your consent" behind a huge wall of "this is how we use your data, take it or leave it"), and for those companies to sell it to data arrgrators to build a complete picture of you, to sell it to anyone with enough cash?
However, what you describe already happens and has for decades, in the offline world. Tons of personal info, like real estate and voter records, are already public in many jurisdictions anyway. Insurance companies, credit card companies, phone companies, and everyone else all take this stuff and spam the hell out of everyone.
The movie folks now know that I am very interested in this movie. They can choose to target me for a small coupon advertisement, knowing that I will likely claim it and consider it a win.
Simultaneously they can target people on FB that they think are Batman fans (but who have not signed up for their email list) with a more generous coupon.
So while I am seeing advertisements for relevant products, I may be seeing less-generous offers than I would see in a world without tracking.
Need creation is not a benefit.
I'd rather get random, un-targeted, irrelevant, and even annoying ads from companies that have no idea who I am or that they're even advertising to me.
Less creepy that way, also more likely to result in the serendipity of learning new things outside my filter bubble/what the algorithm predicts for me.
Sometimes in order to find the real signal, you have to accept that a lot of noise comes with it.
The last thing in the world I want to do is pay real money to put ads in front of someone that has no intention of buying and no interest in becoming a customer.
It's bad enough that the primary platforms are making the process more and more of a black box where they control every aspect of the marketing process.
What's no longer acceptable according to GDPR is forcing this onto people unknowingly, or by packaging it into a "take it or leave it" package of fine print.
Just 30%? Pfft. Just wait until they become a customer, and then advertise to them. Bingo, 100% conversion success. Best way to do it? Advert on the conversion page. Customer pays money, you show them an advert, 100% relationship between the advert and conversions, sterling job, and you cut out the middleman because it doesn't need a third party agency to do the placement, so it's affordable even after your fees.
If I didn't have to give an email, I'd probably still use those services, reddit comes to mind, i'm only signing up because i'm interested in the first place.
Having my email and spamming me...sorry, remarketing to me, makes no difference to my retention because I never see it. Providing a quality service I care to use is why I stay with things.
Plus, hey.com (from DHH) has seen huge growth without any sort of re-marketing or spy pixels (as he says)
It's definitely not the whole story, but this is a pretty well-studied thing in psychology. A similar trick in social engineering is to ask your mark for a small favor, which will make them more likely to do you a bigger favor later. I've read about it many times, but can't for the life of me remember what it was called...
From a google search:
Theses 6 principles are reciprocity, consistency, social proof, liking, authority, and scarcity.
Today's XKCD was basically about this (in the Alt text)
This is one of the reasons I stopped giving out my primary email address for user signups. I use a service called Blur which allows for unlimited "masked" emails to be created, allowing me to give companies read-only email addresses. In the four years I've had it I have created 378 email addresses. If I'm including the email addresses that I've already deleted, the list gets to 400.
Marketing and the 3000 spam messages I get per month made me do this. It does not have to be this way, but as long as corporations can play fast and loose with my email address I will make sure they never get a real one to begin with.
Edit: Want to add here that I am not in any way sponsored by that company, I've just been using them for years now and think their prices are reasonable.
I've done this for more than a decade.
Everything before the UUID domain is just the name of the service, so something like hackernews@e913ff00...xyz. If someone sells out my email address, I can instantly burn it by just adding a sieve rule since they’re all unique. I even know who sold it based on what name I picked before the @ symbol. This has been working out pretty well for me so far.
1. What’s the length of the UUID that you use?
2. Haven’t you encountered forms that have shorter email address length limits?
3. Also, wouldn’t such a domain be seen as a spammer/scammer when machine learning starts taking over signup/registration systems?
2. Nope! I would be surprised to see that nowadays.
3. For sending mail, yes absolutely. Receiving mail is a different story though, it seems like most systems do not care.
To know which provider it was (in case I later get spam from somewhere else), I keep a text-file + email myself any time a new forwarder is set up, so this way I can always look up which service it was.
This way, I was able to spot a leak at box.com and maybe a couple of other places, before it was even announced.
There are times when removing email is best, and others where data or business vertical necessitates something different. You still need to verify with your users and do the appropriate testing / analysis / user research.
Your website should allow "demo mode" to make me see what's the value it provides. No way I will go through registration on random pages that fail to immediately make me crave to use them. I ain't got time or will for that.
2) If you want to follow advice of the article, I'd rather not remove email field, but as I wrote before, allow demo mode, and at next step when actually registering the user, put a clear one paragraph sentence saying that you'll not be spamming me with your engagement stuff.
3) If you ask for email, absolutely verify it. There are way too many people who subscribe to all kinds of services using someone else's email. See this thread:
I just checked and I have a bit over 110 domains in my spam list that look like they were added over the years because of this. So I may have under-estimated how often I "blacklist" businesses like this.
Sadly though, it appears all these temporary mail domains are in some central list, that data harvesters use to deny access. It's almost impossible to sign up for forum accounts with these. So it's impossible to download files from vinylengine.com unless you allow them to spam you.
One is my actual address that human people who I know have.
The other – referred to as “the sluice” – is everything else. I don’t really care what goes in it. A rule marks it as read as soon as it hits my inbox.
Simple but massively effective. I used to get stressed about spam but now I don’t give a crap.
Edit: also, the sluice mailbox is 2020@mydomain. When it does finally become too much of a cesspit, I'll just kill it and create 2023@...
With MakePostSell [1] a customer may add products to their shopping cart and interact with a shop as if they are logged in, but at the point of sale / checkout, we ask them to verify their email.
Instead, give them a searchable FAQ or a wiki and an email address to write to for support. Connect it to a ticket system that autoresponds with a ticket number.
(then you added your second paragraph while I was writing this.)
Eg. help3a2fg@company.com
The support team can figure out which account or anonymous cookie saw that email address.
It's a handy way to still communicate directly with users while holding as little PII data as possible about them.
https://blog.instapaper.com/post/2318776738
It's interesting to think that times may have changed and that people are hesitant to give out their email addresses anymore, but you are giving up some real benefits by leaving it out.
In my view, for most applications, the upside is not really worth that downside. It got me thinking though, are there any clever solutions to do password reset without an email / social media account login / etc? Does anyone know of any good ones?
In fact, the more I think about it, there's a paper I saw that can identify users solely by their mouse movements. If you maintained that kind of fingerprinting in game, you could simply ask the user to play a few rounds then offer to reset if they're from a typical IP address. Might work well for this particular website.
This gets a little more tricky if you have an unexpired session but want to be able to change your password (which likely requires knowing the existing password), but a request from this logged in session to reset your password should be trustable (unless your "friends" have also stolen your unlocked device).
Similarly, if one or more of your "friends" requests a token / password reset of your acccount, the site should highlight that in a banner on every page you visit, to potentially give you warning to find better friends. (The process for replacing a friend on the site should probably require re-entering your password too, to stop someone that's hijacked your session from picking three sock puppet accounts as your new friends, and resetting your password that way).
Presumably if only one "friend" defects and attempts to reset your account you will be notified by the other two friends sending you unrequested reset tokens out of the blue
3 is kinda an arbitrary number, chosen to strike a balance between security and convenience. It was decided that getting 3 people to collude to erode the trust of the community was harder than intercepting an email so the solution was accepted as adding some additional amount of security.
Honestly a bigger flaw in this scheme is if one or more of your friends is no longer active or has forgotten their own password and cyclically is relying on you for backup. You can hedge against this a bit by adding more backups and requiring only some critical mass of tokens, but this does also increase the attack surface.
Assumes: - people are less likely to lose their PGP key, than random password to a random website. - people have PGP keys - PGP key doesn't contain email address (it does).
Anyway, it would be reliable, and it doesn't need giving third party online service access to all your online accounts.
Another possibility is requiring a payment with a payment method they’ve used before and then credit their account with the amount. Forcing 3D secure on that transaction should cut down on fraudulent take overs; or at least shift the liability from you, somewhat.
If you have an app, you can also allow them to authorize the password reset from the app on a computer (or vice versa).
Lastly, you could just not have a password to forget. :)
Login via Google / Facebook / Whatever is sometimes helpful, but it usually results in SPAM. For example, I logged into Redfin through Google and they immediately started spamming me.
Other times, when I login through Facebook and disable sharing my email, the site that I'm trying to log into has a "mystery error" because the concept of not sharing my email address never occurred to whoever wrote the integration.
Most of the time, I just use a unique email address with each site. My domain has a catch-all email address, so when someone starts spamming it, I know who did it.
Instead of [everything]@example.com, I set up [everything]@yo.example.com. Discovering subdomains is much harder and the one time I encountered a form that didn't like a subdomain, I just made a forwarding address on my main domain.
Using Fastmail's rules, I have a setup where every message arriving to @yo.example.com gets shunted into a folder unless there's a different rule putting it somewhere else.
Is this because you don’t publish MX records in DNS for the subdomains and the default setting on the main domain is to accept only specific addresses (and reject catch-all addresses)?
The spam I've received on my catchall is either based on previously breached sites (which I once signed up to) or to very common mailbox names (e.g. postmaster@, info@). I just add those to an auto-reject postfix filter based on the intended recipient, which keeps my inbox very clean.
Gmail handles the rest.
Edit: I do like the recommendation of showing them the email they're waiting to validate so that they can see typos.
It's an API where you give it the user's email address and they tell you if they think it's real or not. I have it such that if it's not risky I bypass the "verify email" step.
Not sure how long the API will be around and it's pretty slow but for now at least it's free.
And you really need to be able to recover account access for a paid subscription. It's probably also reasonable to assume that if someone is going to give you a credit card number and address, they're probably OK with giving you an email.
It will give them the cookie again if they re-visit from any IP they've previously used.
It also re-gives them the cookie if they try to pay again with the same credit card.
Support just tells people to try to resubscribe if their subscription has 'vanished' - but it seems to happen to very few customers.
You're giving anyone on CGNat or even the same coffee shop access to your customers account.
On the other hand also elderly are or could be moving in this directions. E.g. my mother getting an iPad and apps like Whatsapp have completly changed "computers" for her. From, being mostly a chore that you have to use, to something really useful for getting information and to stay in touch with relatives in other countries
Anecdotally, i ve observed the opposite too. I don't require email to sign up, but there is an email field further down in the form, and yet 95% people DO enter an email that looks valid, and not just garbage. That said, only 5% clicked on the email verification link , presumably because they don't have to
If you need me to register in order to purchase boots, then something is not ok.