-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
On 02/11/2014 08:26 PM, Zack Weinberg wrote:
> regarding
>
https://blog.mozilla.org/advancingcontent/2014/02/11/publisher-transformation-with-users-at-the-center/
>
>
:
>
>> Directory Tiles will instead suggest pre-packaged content for
>> first-time users. Some of these tile placements will be from
>> the Mozilla ecosystem, some will be popular websites in a given
>> geographic location, and some will be sponsored content from
>> hand-picked partners to help support Mozilla’s pursuit of our
>> mission. The sponsored tiles will be clearly labeled as such,
>> while still leading to content we think users will enjoy.
>
> I get why this might seem like a good idea if you don't think about
> it very hard, but it is a profoundly bad idea ...
....
[and then me again in a different message]
> I have hit my personal limit for responses to a single thread per
> day and will now shut up again until tomorrow evening.
I apologize for letting this discussion drop for two weeks. As you
might imagine, I do have actual work I have to do. :-) Also because
of actual work, I was not able to attend the "town hall" on this
topic. I regret any rehashing of stuff that was covered there.
Because of the delay, I'm going to write one big response to everyone
and post it once. If you asked me a direct question which I failed to
address, it's because I missed it in this voluminous thread; please
just go ahead and ask again. Please don't cc: me on replies to the
list, but if you want to take some subthread off-list that's fine.
So, this whole thing started because of a poorly-communicated plan to
have some "default" tiles on the new tab page for brand new users. I
concur with what Christian Heilmann said about the poor communication,
and have nothing further to add. The default tiles, in themselves,
are a fine idea. I wasn't sure what I thought of it initially, but on
reflection, having some signposts for people who may be completely new
to the Internets, or even computers in general -- here's how you can
get an email address, here's some options for online socializing, what
to do if you want your own website, what to do if you want to learn to
code, did you know this very browser is made by Viewers Like You and
you can help? That would be a Good Thing.
Where I -- apparently, still -- part company with the people pushing
this plan is: I think it would be a CATASTROPHICALLY BAD IDEA, both in
the short and the long run, to allow anyone to give us money in
exchange for placement on this screen. Let's start with the short-run
reasons. The most basic and immediate reason this is a bad idea is
that people will be angry at us over it; it will cost us more in
goodwill than we could possibly hope to get out of it in cash. In
fact, people are *already* angry with us just for suggesting it. We
might still be able to restore those people's good opinion of us if we
recant. Mitchell's blog post about it last week did not qualify as
recanting; she was quite clear that the possibility of paid placement
remains on the table.
I said that last time around, and nobody got it. I got lots of
responses drawing a moral equivalence between this and the existing
product placement in the search box and the bookmarks. The funny
thing is I agree with that moral equivalence -- but I think those are
*also* bad things, but bad things we are stuck with in the short run.
More on that point below. I also got responses which amounted to
"well, why should we listen to these people? Look at what we can get
out of this!" ("What" basically amounts to hypothetical reduced
dependence on the Google deal, afaict.) This is also a fair reaction;
people get mad at us all the time and often we decide that we don't
care. But in this case, we *should* care.
The first and most nakedly mercenary reason we should care is because,
unlike the search box, the unpopulated new-tab page is a
high-visibility but low-frequency context. The search box drives
traffic to Google every day from (to first order) every single one of
our users, so Google is prepared to give us a whole lot of money to
continue to be the default option in there; but you have to go out of
your way to think about it as something bought and paid for, and it
also happens to be the default that many of our users would pick
themselves, so nobody (to first order) gets angry with us for taking
that money.
The unpopulated new-tab page, in contrast, will be seen order of ten
times by each new user. There are anywhere between three and nine
slots on that page depending on how big your device is. That's not
enough ad impressions for anyone to give us very much money for. But
it is an *extremely* visible context, and a context in which most
people's immediate assumption is going to be that those slots are
bought and paid for. In fact, even if we *don't* take money for
placement on this screen, we may never be able to *convince* people we
don't! Thus, the potential loss of goodwill far outweighs the amount
of money we could plausibly hope to bring in.
Selling slots on the new-tab screen would also hamper our ability to
make the new-user experience *even better* in the future. Concrete
example: right now, the best "you probably want an email address,
huh?" option for someone completely new to the tubes is one of the big
webmail providers: they're reliable, they've got good spam filters,
they've got people whose job is to worry about the servers getting
cracked. But they also have significant drawbacks: many of them
suffer from UI designed without any idea of how email *ought* to be
used (probably the designers are too young to know), they don't
support PGP, all your email is stored in plaintext in a gigantic
database under the provider's control, etc. In the medium term,
something better may come along. If what's on the new-tab screen is
entirely up to us, we can just change it. If it's paid for, we have
to get out of a contract somehow. This already does come up with the
search box -- was it two or three years ago that Google wanted to be
the default in *all* locales, overriding localizers' preferences? I
don't recall how that turned out.
Moving on to the longer-term picture: because hypothetical advertisers
would be paying for not very many impressions per user, we would
naturally come under pressure to stop limiting the impressions to the
"until there is a populated history" first-few-runs setting. They'd
give us more money! I think this is what scares the commentariat
most: not the highly limited thing that has been proposed, but the
much more aggressive thing it could become. This is a *rational*
fear, and a scenario we should bend over backward to avoid.
Advertising inherently intrudes on people's attention; people learn to
ignore it; the advertisers respond by making the ads bigger, brighter,
flashier, and more carefully targeted; the people still learn to
ignore it, because brains are really good at ignoring things; the
advertisers escalate again, because what else are they going to do?
Fast forward a few cycles and you get the sites we've all seen where
the above-the-fold display is a logo, an article title, and
advertising; you have to scroll down to get the content you wanted.
And there's probably an "interstitial" pop-over ad, too. And *still*
nobody clicks on the ads, but now they are irritated and may close the
window. It doesn't just happen online; the billboards on I-80 in San
Francisco have gotten progressively bigger, brighter, and more
obnoxious over the past 25 years; you can measure the steady slow
decline of the newspaper industry by the proportion of the Sunday
edition that is useless supermarket coupons and suchlike.
I understand that part of why we're talking about new forms of product
placement is because we're organizationally uncomfortable with
depending so much on a single source of revenue, namely the Google
deal. But we've gotten lucky with that: it brings in lots of money,
but it doesn't bring us under pressure to escalate. (At least, not
that I know of. Perhaps the sales team has been quietly declining
propositions to bundle the Ask Toolbar in the default download for
years now. ;-) If we start soliciting deals from a wider market, for
a wider range of options, we'll be opening ourselves up to pressure to
escalate, to become ever more ad-driven, and ultimately to start doing
the sorts of clearly unethical things that everyone here agrees we
*shouldn't* do, but that otherwise-respectable companies keep being
caught doing. Swapping out ads in content for our own ads, adding ads
to pages that didn't have them in the first place, feeding user
tracking data to third parties, that sort of thing.
It is my considered opinion that the only way to be *sure* we don't
find ourselves over that kind of barrel in the future is not to do any
more business with the advertising industry than we already do, and to
make it an explicit long-term goal to phase out our existing
involvement. Yes, we should diversify our revenue stream; no, we
should not diversify within the advertising sector. Rather, we should
pursue entirely different sources of income. I can think of several
plausible options just off the top of my head. Revenue sharing with
carriers shipping Firefox OS phones is probably already being
negotiated, and could ramp up to big bucks more quickly than anything
else. Transaction fees on some sort of marketplace (for webapps?) is
the next most obvious option. Partnering with an online payment
provider, to allow in-browser identity to be a purchasing principal.
Merch and donations from users probably won't scale to big bucks, but
sponsorship deals from huge companies might. (It wouldn't be my
favorite thing, but I'd be okay with a line of type at the bottom of
the new-tab page saying something like "Development of Firefox 42 was
funded by [Weyland-Yuutani], the [Umbrella Corporation], [SPECTRE],
and [many others]. Did you know that [you can help]?" where square
brackets indicate hyperlinks. This strikes me as much less likely to
bring us under escalation pressure (because the organizations
mentioned are not paying for the mention, but to directly fund
development) and it's a thing that people have seen in other context
and understand, so it is also less likely to make them angry with us.
There should probably be an X to make it go away.)
To be clear, I can imagine ways in which the above suggestions for
alternative revenue would wind up putting us in an awkward position
with to our principles and/or our duty to our users, but I think they
are all relatively unlikely. This stands in stark contrast to
doubling down on the ad revenue, which I think is practically
*certain* to put us in an awkward position later if not sooner.
This also means that I am not a fan of new or proposed Web-platform
features whose primary function is to make advertising "less terrible"
and/or "less privacy-invasive". For all the same reasons that we
might find ourselves under pressure to escalate our *own* use of
advertising as a revenue stream, we would also be opening ourselves up
to "regulatory capture", as it were, if we added these kinds of
features. Gerv brought up a hypothetical situation in which each
browser installation, perhaps in concert with Mozilla-controlled
servers dedicated to the purpose, selects ads from a repertoire and
displays the most plausibly relevant ones to the user, based on some
sort of personalization algorithm carried out client-side. This would
in fact be a short-run improvement on the status quo, where the ad
servers make guesses about what's relevant and vacuum up as much data
as they can in order to attempt to personalize, and there's no one
involved who particularly cares about making sure the ads are not
malware. (I tell people to run aggressive ad blockers *just* because
of ad-sourced malware, nowadays.)
The catch, though, is that we would then have features directly
targeted at the ad industry, so their use cases would drive future
development of those features. Even with the best of intentions, I
seriously doubt we could succeed in keeping a privacy wall intact in
the long term; there are just too many ways that the advertisers could
potentially figure out what our ad-selection algorithm is doing, and
the more features we added, the easier it would get.
Even if this didn't happen, the features themselves would be sucking
up developer time and attention that would, I think, better be turned
to new platform features that offer Web companies *alternatives* to
in-site advertising. The obvious thing is, again, in-browser identity
as payment credential, and making it easier to put up a site that
takes payments. I haven't thought about this as much, but I'm sure
there are more things we could do.
Thank you all again for your attention.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=VElT
-----END PGP SIGNATURE-----