Speculative ad formats for the post-newsfeed world

21.31, Monday 11 Jan 2021 Link to this post

So I was on a call recently, talking about people abandoning the global timelines like Twitter and Facebook, and instead hunkering down in private Dunbar-number-scale virtual neighbourhoods, such as Discord servers etc, as discussed last week.

And, on this call, because of the fact that these virtual private neighbourhoods don’t collect personal data, and therefore can’t target ads, I mentioned something like: well advertising will have to change.

And they said, how?

And because I’ve been thinking about this for a little while, and because while I like advertising (generally speaking it’s a straight-up way to find out about new things), the capability of microtargeting is a societal risk like having water pipes made out of lead, I had in mind a few speculative alternatives.

All of which is to say: here are three quick ideas about at-scale advertising into zero personal data private communities.


Target communities, not individuals

In the old days, adverts would target publications by interest and demographic.

Today, newsletters and podcasts are both growing: A fat middle of ongoing publications around specific interests and personalities, building audiences – engaged audiences who are demonstrably investing time (podcasts) and money (Substack newsletters) in continuing their engagement.

So who is building the ad machine to target these communities en masse?

At the moment, if I want to spend $100k advertising on Facebook, I can. But how can I, at scale, choose a collection of communities to advertise into, with enough volume and diversity that it’s possible to run auto-optimising algorithms?

Could each publication submit a gestalt view of their audience (demographics, interest, readiness to purchase, etc), and targeting is a matter of choosing the right collection of audiences, but not targeting within the audiences?

While it’s not desirable (or even possible) to collect data on individual clickthrough, maybe there’s a startup in collecting data on conversion stats for different communities?

What’s the Audit Bureau of Circulations of pro-am online media?

Telepresence adjacency

The traditional way ads work is that you find a place where people are hanging out, with a relevant sentiment, and you place a message next to it.

Adjacency works well on a newsfeed: you get people in the mood to respond and tap and act, then you put an ad in the way and… people respond and tap and act.

But if everyone is spending their time in private Slack channels, or private Discord servers, not scrolling but actually socialising, then what does adjacency look like?

For me, the answer is that people will be hanging out in multiple spaces. That’s another difference when the global timelines go away. It’s not 100% Facebook, it’s ~10% each in a dozen different spaces.

So the future will include a variety of

  • long-running home spaces (where we socialise with friends or interest cohorts)
  • persistent but touristed spaces
  • temporary spaces, like events.

Imagine MakeSpace teams up with Tate Modern to make a persistent museum space with art to experience, explore, and learn about. People will visit that space and hang out with their friends – why not? It’s more fun to chat when there’s stuff around you to look at and play with, and you might run into people too.

c.f. Monterey Aquarium sets up shop in Animal Crossing.

Or there are ephemeral spaces that pull huge crowds. There are temporary spaces that people travel to, e.g. in-game concerts. (Simply streaming concerts doesn’t count. You have to have the social space otherwise there’s no interaction adjacency going on.)

So what are the billboards for virtual spaces, and how do you price them?

There’s Bidstack in this space, which I am super into as a concept. They allow advertisers to buy spots in-game: trackside banners, cityscape billboards, pitchside LED boards and other native spaces within the virtual world.

But virtual billboards in virtual events are a blunt instrument.

Think of the real world. Where are the in-game stands doing giveaways, the leaflets, the limited edition t-shirts, etc. There can be equivalents of all of these.

Industrial-scale micro-influencers

The scenario is this: everyone spends their time in private Slack groups and private Discord servers. There’s no data collected, and no ads. How am I, as a business, supposed to let people know about my new shop, or sell these new shoes, or get people only my subscription fashion business, or drive signups for my new video tool, and so on.

The answer is that you go door-to-door.

We have the equivalent of door-to-door sales in social media already: influencers. Somebody has an expertise, and an audience, and they plug product in their feed for money.

But instead of 1 or 2, or even a half dozen, how can a brand have 100,000 influencers?

Look, people love to be associated with the brands they love.

Instead of making influencers endorse brands on their public feed, and turning them into media, treat them as ambassadors. Have a scheme where people can sign up to get (say) advance versions of unique Heinz drops - I guarantee there will be people who are into exclusive flavours of spaghetti in a can, of all walks of life - and maybe a discount, and all they have to do is ensure that a certain number of people from their allotted community Discords tap on a certain link.

The cost around this is prohibitive right now, which is why it can’t happen, and you would need a way to prove that certain clicks came from certain communities – even if you didn’t know the individual who clicked through.

But, with work, I bet you could industrialise the process, and appropriate the language of ambassadors and drops from high margin brands, making it available as a standard marketing product.

(Anyone who grew up in bars in the 90s will remember encountering “cigarette girls.” Here’s a write-up from someone who did the job. I hadn’t remembered cigarette girls until just now. Blimey, let’s not do that again.)

BONUS: Industrialised influencers who are bots. Chatbots are getting pretty good. I read about someone recently who had configured a chatbot to act a bit like his long-distance girlfriend, and was way more into the bot. So alongside the ambassadors program, I would suggest making chatbots that can be added to these private social spaces, and while they flirt and socialise and tell jokes, they also every so often suggest stuff to buy.


You fish where the fish are.

As a business - an advertiser - there is an upcoming need to reach people in spaces that currently do not allow for gathering personal data, targeting ads, or collecting conversion metrics.

Maybe tools to collect said data can be built. But that would be very much against the tide of GDPR and consumer sentiment.

Instead: let’s think about how to advertise with the minimum of data, not the maximum.

There are a bunch of startups embedded in the notes above and, although this is unlikely, maybe one of them has a path to a format like the banner ad, or a clever sales model like AdSense.

But if I were an established online ad broker, I would be working wildly to figure this out now, because otherwise one of those startups might get there first.