Comment by alanlammiman | Hacker News Reader

Comment by alanlammiman | Hacker News Reader

The problem isn't just Apple's action on IDFA. It's that Facebook seems to be so poorly managed on some fronts that its reactions, rather than mitigating problems, has caused further harm. For example - in a rushed effort to get their privacy issues in order, they are deactivating the live facebook integrations of customers based on cursory/mistaken/possibly machine-based readings of their privacy policies.

They did that to us yesterday: https://shared-crater-f3a.notion.site/Facebook-is-Breaking-A...

We spend hundreds of thousands of dollars on facebook app ads per month. We can deal with IDFA giving way to more aggregated attribution (we don't want to track individuals - we just want to measure if the ads we paid for led to sales). But facebook breaking our app in production because they can't be bothered doing their job properly is very serious. It can't be solved by reducing ad spend, only by removing their SDK from our app.

If this is also happening to many other developers right now, that, more than the Q4 results or the IDFA issue itself, could be causing the drop in the share price.

In fact, if you look at the Q4 results, the earnings miss was more because of growth in G&A (which grew by 3 percentage-points of revenue if I'm not mistaken) than because of a top-line slowdown. And if you read the comments as to what made G&A grow, it's 'legal costs'.

> we don't want to track individuals - we just want to measure if the ads we paid for led to sales

well. if your goal isn't tracking individuals, then why are you attaching unique ID's (in cookies) to track individuals on your website?

And I'm not talking about third-party cookies disguised as first-party.

logglytrackingsession (lifetime: session)

notion_experiment_device_id (lifetime: 1 year)

Both are unique to a specific user and are used to identify a single individual. The first one is short-lived, but obviously meant for tracking and the second one can be used for tracking, identifies a single individual and is long-lived.

edit: turning off my adblocker, some more appear.

_ga, _ga_4GMCF7E1GC, intercom-id-gpfdrxfd, notion_browser_id, amp_af43d4

none of these are listed or explained in your privacy policy.[1]

[1] - https://shared-crater-f3a.notion.site/Sticky-Privacy-Policy-...

One could choose a better host than Notion, but that's Notion's fault.

Yes, agree, we wanted to try it out for this because we like it for work and it was quick to write up and publish and we thought it might be more flexible for writing up documentation, linking, etc. But we are not satisfied either: Very slow to load, formatting is fiddly and imperfect. We will have to use something else.

Yes like the peer reply said that's Notion, not us. But good point, another thing to keep in mind if you try to use the "Share to the web" notion feature.

the _ga stuff isn't Notion, but the rest is, yes.

My real point is, if you use a service to provide your own service, you give them your blessing to do whatever they want with your brand. This includes facebook and their tracking scripts.

Thus we need to audit what our service providers are doing and limit their impact once we've completed the evaluation, making sure they don't alter the deal later.

Also tracking ad conversions is as simple as using a unique parameter per campaign, when buying the add. Just append `?campaign=facebook_campaign-name_202202` to your link and that's enough to measure the ads effectiveness. No need to attach unique ID's to users, sessions etc... Aggregates keep the users anonymous and give you enough actionable insight.

While I agree with the sentiment of your first point, in practice if each small business were expected to audit every one of their inputs, it would be hard to get any business done.

As for your second point, yes, that is precisely what I referred to in my original post - that we could deal with the end of IDFA / that we just want to make sure ads are effective - SKAN which provides aggregated statistics is mostly ok for us.

> in practice if each small business were expected to audit every one of their inputs, it would be hard to get any business done.

While I agree, that a full audit would be difficult for smaller operations, it took me 2-5 minutes to do a quick check on what is being stored client-side and to come to a logical conclusion if individual users are being tracked or not.

It's a decision, one that you can (probably) make. For me, in the EU, it's no longer a choice and I personally think regulation(GDPR) was needed because, without it, no one took user-privacy seriously.

> As for your second point, yes, that is precisely what I referred to in my original post - that we could deal with the end of IDFA / that we just want to make sure ads are effective - SKAN which provides aggregated statistics is mostly ok for us.

GDPR would also apply here. If there's an option to process less data and achieve a similar result, one should use that option and using a more invasive method(identifying individuals) for tracking would be illegal. It's called "the data minimisation principle"

"we don't want to track individuals" --- right, but each ad FB places has a lower conversion rate if they don't know as much about the user, so while your goal may not be tracking per se, tracking does get you the same sales in fewer placements, and Facebook can sell those saved placements to other buyers

the fact that they're cracking down on partner privacy in a hamfisted way surely doesn't help matters but I can't see how angry devs are driving the share price down

I agree with your first point in the sense that having individualised targeting data necessarily improves advertising by some amount vs. merely contextual and aggregate data. For our specific case I'm not convinced that amount is as big as some people imagine it is, but I haven't found a good way to test that precisely given that facebook doesn't really have a "just give me a representative sample of your audience in this region" targeting feature (wish it did) to compare with the targeted alternatives. In our case, app ads that now use SKAN (the Apple solution that provides aggregated attribution) seem to be working well enough - that was my point.

Regarding your second point, yes, I agree - it's not obvious. Assuming this did indeed affect many developers like us and that it happened to everyone at the same time (which may not be the case - that it happened to us and that it coincided with their earnings report may have just been a coincidence - I haven't seen a mass outcry on twitter or anything), I was wondering whether it might be hedge funds that buy/track aggregated ad spend or attribution data, perhaps from MMPs or media buying agencies. I know they buy app download data from the likes of App Annie, but don't know if equivalent data is available and timely for ad spend. In any case, my point is more that this is illustrative of how they make bad situations worse for themselves.

Reply on news.ycombinator.com