Advertisement Midmortem

Let's talk about ads! (this was written as a Kickstarter update for Lancer Tactics)

As mentioned previously, I've never participated in an ad campaign before so this is an experiment for me — and my scientist training has taught me to digest experiments by writing and sharing, so here we are! I hope to pull back the curtain some on how your data is being used across platforms & maybe come to better understand the money engine that powers most of the internet.

This is a hunky update, so feel free to skip to the discussion at the end to see us agonize over the morality of participating in society.

The data so far

We've been running ads on Facebook and Instagram since Monday the 13th. The campaign was in its middle plateau phase at about $80k and was averaging ~$1k in pledges per day. There's always a bump in the final 24 hrs, but even accounting for that I think we would have only crested $100k by the end of the campaign; we've already far outstripped the capacity of my personal connections and I don't have any kind of press contacts planned (because I have an incredible amount of social anxiety about cold-calling strangers and am already whelmed by managing this campaign).

Since starting the ads, we've seen about $2k in pledges per day from directly from ads (tracked from the URL that folks use to reach the campaign). The rate of non-ad pledges has also increased from $1k/day to about $3k/day (Kickstarter itself pushing it more due to the increase of traffic? I haven't been able to track down the source of these). These have gotten us to a current project total of $110k, compared to the ~$89k where we would probably be without running ads. The backer who claimed the Xiaoli homebrew reward also came to the project via ads; it's unlikely that tier would be filled if not for this ad campaign.

Figure 1. Pledges from ad links are shown in red, all other pledges are shown in blue. Data from March 1 to 6 condensed into a single point because we hadn't partnered with BK yet.

Figure 2. Backerkit marketing dashboard. ROAS means "return on ad-spend"; it's how many dollars we've gotten back for every dollar we've spend on ads.

This sounds great! But it carries with it a hidden cost: participating in ads means the funding level of the campaign no longer 1:1 represents the the money we'll have available to spend on the project. We originally set each goal to the actual cost of implementing that feature, but now every pledge that comes in through advertising has a chunk shaved off — it's almost halved!

For example, this first week of ads has generated about $13k — but we will need to pay for the ads that got us to that point + Backerkit's 15% commission. With a time-of-writing campaign total of about $109k, the post-payback total is $103k and that gap will only widen the longer we do ads. 

Figure 3. Calculation showing how a $13,000 increase in pledges from ads ends up only adding about $6,500 to the project.

And that's before the cost of fulfilling $13k more's worth of pledges; it means we'll need more stickers, coins, postage, etc. Fortunately we did a good job picking rewards and those are flat costs so we're not in danger of breaking the bank, but it's worth remembering.

Devil's wager (facebook is the devil in this metaphor)

So those are our choices as I currently see them:

  • we could end our ads and happily coast along to the end of campaign with what we have, or 

  • participate in ads that will almost certainly bring us to the stretch goal of $150k — but if we don't end up overshooting it by enough we'll end up with the obligation of a 100% port without the full $150k to pull it off. Running ads brings our "true" final stretch goal from $150k up to about $170k!

Backerkit's current projections put us at about $160k ± $30k by the end, so falling in that awkward window is likely. We'd be able to find a way to make it work and the project would not be in jeopardy, but it's not ideal — we would have to dip into our personal savings in the hope that later sales of the game would make it back up.

Mark and I discussed this and, for now, have decided to continue with the ads for as long as they maintain a good ROAS. Honestly, we just really want to take a crack at 100% port* to get the game into a more complete state. We'll both be able to absorb the risk if that's what it comes to. So — here's hoping! The game is on! We might actually make it to that final stretch goal!

*or a year of full-time work on it, whichever comes first

Targeting, or, Why Is Olive Asking Me To Fill Out A Survey?

Here's where we get into the parts I still don't understand as well; Backerkit has been handling all of the audience-building stuff, and bless them because I've tried using that interface before and it's gnarly. We used the backers from my previous campaigns and the people who signed up to be notified for the project launch to... somehow build out the audience on Facebook to show the ads to?

Whatever they did, it's working. We've gotten nearly twice the click-through rate as industry average, although it's fallen off after the initial splash. I'm hesitant to try and interpret this as Backerkit has advised me that these are highly context-dependent and project-specific, but it's generally a good sign and means that the money we're spending is more efficient than usual. 

Figure 4. Click-through rates (CTR) and conversion-to-pledges percentages.

As an added bonus, the visibly-gay ads being well-targeted has a protective effect in that they're less likely to be served to people looking to start trouble; moderating the comments was rough for the first few days (I learned some new transphobic dog whistles, cool cool cool) but it wasn't as bad as it could have been and I have a better handle on it now.

Many of you have gotten a message from me about a survey about how you found the campaign; this helps us know what's working (so we don't shovel money into fb's maw for no reason) and I think helps us not serve ads to people who want to make "jokes" about people like me being dead. Though I want to stress you're under no obligation to fill them out if you don't want to share that information! It's helpful info, not essential.

As you can see, the click-throughs are dropping off fast so we're probably going to wind down this first batch of ads as we figure out when the next step to take is going to be — either a new batch or wait until near the end of the campaign when we can pull out that "final 48 hours!" card. If you're curious, here's a list of the 7 ads and their stats. Trans Balor has predictably caused the most engagement, but note that the ads that demonstrate the tactical nature of the game (2, 5, 6) have the most efficient conversion percentages.

Figure 5. Each ad, the number of times it has been served (impressions), clicked, and how many pledges to the campaign it has ultimately gathered.


Ads are an engine to turn money into more money! It suddenly makes sense how a company like Google can run on ad revenue; as long as an advertiser stays comfortably above their minimum ROAS, any magnitude of money spent is basically free because you make more than that in return! I can so clearly see the path of how someone could scale up from here and end up throwing a quarter million to these tech companies without batting an eye. ROAS is better when you know more about your users, so this engine is incentivized to do all of the privacy-invasive stuff we've heard about.

Let's take a second to step back. I'm goddamn considering paying facebook $13,000 — an amount of money comparable to my last entire campaign — because that's what the numbers on a spreadsheet come out to! It feels real bad to have such a large percentage of backer funds get diverted into a platform that helps overthrow democracies!

Is it better to turn your back on this advertising monster because you don't believe in its values? Or is this just what participation in society/this industry looks like? I've been asked a handful of times — by sealions, granted — what being gay has to do with doing giant robot crimes. My canned answer is that we want to make a game that's "culturally queer, with those values shining through wherever possible." Does that begin with saying no to participating in this kind of system?  Is it hypocritical to pay facebook to put up a trans flag? There's gonna be some ad there — why not have it be pride instead of some war game that doesn't concern itself with all this? This is the battlefield we're on, let's take up space! // This is the trap they set, capitalism transmutes all criticism to itself into more fuel for its own fire.

Mark and I discussed it, and our answer to this dissonance is that we gotta pick our battles. We have limited space and resources. Our core goals are to make development a good experience for all contributors (pay fairly and promptly as possible, fit work schedules around our outside-work lives), to make a mechanically and visually polished game, and to infuse what narrative space we have available with all the queerness and anticolonialism we can muster. Trying to retain some kind of moral purity around the advertising of this initial fundraiser is related to, but not essential, to those goals. The ends don't justify any means, but they can outweigh some on a case-by-case basis. We just gotta trust ourselves to make art that's worth it. (we also don't judge other people for participating in ads or not; I trust others to make their own calls on this front)

Thank you again to Backerkit; they've been very helpful in guiding us through this process, have listened to my concerns, are down for us to share all these stats, and have been proactive about scaling ad-spend as makes sense for the success of the overall project. Personally, they also helped me narrowly avoid getting phished due to me being too exhausted from moderating the fb comments to notice the trap.

Whew. Thanks for reading, thanks for supporting. I may missed or gotten some things wrong because I'm still pretty new to this, but I hope it was an interesting dive.

Olive  🌺


Play Anthropocene interview on Lancer Tactics


Lancer Tactics Kickstarter Postmortem