Jump to content
Sign in to follow this  
iamlost

Europe's Gdpr: How Best To Adapt; More Questions Than Answers

Recommended Posts

Just read Pierre's (eKstreme) latest endeavour marketing letter, GDPR, ePrivacy, and web analytics, LinkedIn, 13-November-2017, and it reminded me that major changes are afoot. And that there is little substantive discussion between the legalese of the regulations and vague proposed commercial 'solutions'.

As a consumer transversing the web I understand the increasing use of javascript inhibiters, i.e. NoScript, ad/tracker blockers, i.e. Privacy Badger, and appreciate the thrust of the EU's GDPR (General Data Protection Regulation) and ePR (ePrivacy).

However... :)

As a SME aka sole proprietor webdev owner of sites selling direct ad space and affiliate pre-sell traffic who has spent the past decade improving contextual delivery of both site informational content (what brings the visitor) and ad/af presell possibilities (how I pay the bills and, hopefully, adds value for the visitor) wending my way through the prospective requirements is 'take two bottles of aspirin and call the doctor in the morning' time.
Note: never ever take two bottles of aspirin, black humour is not valid medical advice.

There is a push to anonymise data in one form or another right across the board; unfortunately that destroys the advantages of personalisation. And there can be a decided advantage for the visitor as well as a site's bottom line business dreams.

On my sites there are four separate data collection purposes:
1. maintaining site operation.
2. contextual delivery/personalisation of content.
3. affiliate form pre-fill.
4. marketing valuation of referrers, visitors, and site pages/categories.
It's the last three that are most problematic.

'2' is of greatest concern as, currently, it includes visitor 'fingerprinting', which is necessary to identify return visitors (no site registration or sign in).

'3' is less so: an affiliate pre-sell PII sensitive order data form is rendered by client browser as part of a page of my site, all checking/confirmation of input is done client side by script, input form data is sent directly to merchant landing page populating their order form. At no time is any of this form data referred back to my site. Nor retained in page form past tab/window close.

'4' also less so as is pretty much anonymity sets already as that is all that is needed for desired overviews.

General process overview:
* logfiles and visitor data are concurrently shared with dedicated live analytics servers that handle personalisation, contextual delivery requirements.
* after 24 hours the data is encrypted and saved for additional/future analytics interest.
* individual 'fingerprints' and associated site behaviours are stored as hashes for rapid comparison/identification purposes.
Note: no off site sharing or tracking.

Given the above:
* if 'John Smith' asks to be removed but names were never collected so there is no association to a fingerprint?
* if asked to remove data connected to a specific IP and there are two or more fingerprints associated with it how to determine which if any is that of the requester?

Etc. :)

Share this post


Link to post
Share on other sites

Quite. Often you have to ask the question 'er we're not actually collecting anything personal at all!' Google is.

Share this post


Link to post
Share on other sites

My 2 cents on your point 2:
I'm not a lawyer nor have any legal training. I think all the data you collect would pass a privacy test. You mentioned, previously, that you use your own Piwik hosting and just mentioned above there is no sign-in on your site. So what do you have? A fingerprint of a visitor who lives in Nick's Nose Cove, Newfoundland. It could be any of those five people living there. But who is it really? Number 74930712 in your DB.

I've mentioned it before. You don't have a name and address to which you replied your info is better than that because you have a wealth of data on visitor 74930712 for the purpose of marketing etc on your site. Still anonymous it the sense of privacy which that bill is all about (did not read). You may be able to make very educated guesses but could not tell age, sex, gender, race, colour, and native language of that visitor.

I think it is mainly addressed to Googles, Facebooks, Microsofts, and Amazons et al who have names and addresses and more on all these visitors and have the capabilities to data mine this data. I guess it applies to site with sign-ins also. I presume you do not have the resources of the those companies.

Now if a person asks you "to be removed" that person would have to provide specific ID which kind of defeats the purpose of anonymity. Can't say the same of a google type who has the info of that specific person.

Subsitiute Mexico City for Nick's Nose Cove.

Share this post


Link to post
Share on other sites

Thanks for shout iamlost.

 

Nitpicking on something subtle and very important, and too big to dive into here:

 

 

 

There is a push to anonymise data ... that destroys the advantages of personalisation

 

This is a common assertion, and it's wrong. It's just that current approaches to personalization will not work as is (true), but that doesn't mean there aren't alternatives (also true). Just one example: https://medium.com/@subhojit20_27731/no-you-dont-need-personal-data-for-personalization-de9222cff8e4

 

Also, I urge everyone here to read the GDPR's definition of personal data. User identifiers like cookie IDs are included in the definition. IP addresses can be on their own or in combination with other signals. Device fingerprints definitely are. CRM IDs are. Here is a good starter overview:

 

http://privacylawblog.fieldfisher.com/2015/getting-to-know-the-gdpr-part-1-you-may-be-processing-more-personal-information-than-you-think/

 

Also watch the ePR much more closely.

 

[plug type=shameless]And if anyone wants to join the free beta I'm running that helps tell a much cleaner compliance and data governance story, hit me up. Email in the LinkedIn comments, and PM me here too.[/plugpleasedonthurtmemodsyouknowiamaniceguyandwouldneverspam]

 

kthxbai.

Share this post


Link to post
Share on other sites

iamlost:

There is a push to anonymise data ... that destroys the advantages of personalisation

 

eKstreme:

This is a common assertion, and it's wrong. It's just that current approaches to personalization will not work as is (true), but that doesn't mean there aren't alternatives (also true). Just one example: https://medium.com/@subhojit20_27731/no-you-dont-need-personal-data-for-personalization-de9222cff8e4

There are two types of personalisation.

The first one, which every 'alternative' such as your linked example I've read tackles, is a general or aggregated grouping, version of personalisation. It is more persona-isation than personalisation. And I do use it, extensively, for a decade now.

Note: the wisdom of crowds is popular wisdom, which could be considered an oxymoron; yes, I know the book separates wise from irrational crowds/groups - the first, IMO, are unicorns.

 

The first is relatively simple and comparatively simple. The second is not.

 

The second one, is to identify a particular visitor as new or repeat. This one is the kicker.

If you can point me to a method that does not require (1) registration, (2) cookies or similar tags, and (3) is, statistically, both highly sensitive and highly specific, I am most^n interested.

 

It is an intriguing situation: I have no interest in personally identifying a particular visitor yet I have great interest in recognising them. And associating prior site behaviours with them.

Note: refer to The Living Adapting Site, Cre8, October 2015, for a simplified walk through description.

 

It's not that I can't keep doing what I do and, I believe, satisfy a GDPR audit, rather that how I'm doing what I'm doing is not how most/all others are doing and so is not expressly addressed. Once again iamlost: an edge case, an outlier. :)

Yes, my law-type-person is on the job, however, I do like to know outcomes going into meetings.

 

I have just suffered through the China regulation changes (to good result, phew) to go directly (do not pass Go, do pay your law-type person) to this EU one... I'm too bloody old for all this stress; for heaven's sake! I went into webdev to get away from it!

In the spirit of the season: bah humbug! Bah Humbug!! BAH HUMBUG!!!

 

Share this post


Link to post
Share on other sites

If you just want the Boolean new/repeat signal, a cookie is fine. It's not personal - there is no ID! - so that's that.

 

But if you want to have a notion of a user and track them over time to provide real-time personalization, you're now in the realm of the GDPR and ePR.

 

Of course without knowing specifically what you're trying to do, I can't offer (m)any pointers.

 

P

Share this post


Link to post
Share on other sites

Here's a kicker. I went to that link (General Data Protection Regulation) and, on the landing page, Privacy Badger reports 9 potential trackers: 3 which were blocked completely (red), 2 google ones for which it let only the cookie through (yellow), and 5 in the green.

What's that saying about doctors healing?

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this  

×