When I joined Inflexion Studios last year, TikTok was already a hot topic in the gaming industry. But with so much to do, the thought of starting a TikTok account to promote our first game, Nightingale, sat in the background until a few months later.
We officially started our TikTok journey in September 2022 and while we knew it was a promising platform for businesses, we had no idea just how quickly it would overtake the combined performance on every other social media.
I’m not going to claim to be some social media guru or TikTok master, but I want to break down the philosophy and strategy we used in hopes that it helps others looking to get started on the platform.
Background
Before we talk about TikTok, I’m going to provide some context that I think is important as it likely impacted our level of success to some extent.
Nightingale was first officially announced at The Game Awards back in December 2021. Not only does The Game Awards receive millions of impressions, our announcement was also covered in major gaming publications such as VG247, Eurogamer and PC Gamer.
Since then we’ve been in Summer Game Fest 2022, The Game Awards 2022, and featured by basically every major gaming publication, including these magazine covers from PC Gamer and Edge Magazine.
I don’t bring this up to flex but rather to be transparent and show that I took on a project which already had quite a bit of brand awareness. Heck, I think we were already in the “Top 25 Most Wishlisted Games” on Steam by the time we started TikTok.
While it felt like the mass majority of people who commented on our TikToks were seeing us for the first time, I can’t deny that we would see people who had mentioned that they were excited for our game because they saw coverage of us somewhere else.
We have a little over 100 employees at Inflexion Games and consider ourselves closer to the “AA” space (bigger than most indie teams, but not quite AAA level either). Our marketing team at the time was five people and while I was the sole community manager, our wonderful marketing artist did all of the capture and editing of our TikToks while the rest of the team helped with editing my proposed scripts and providing feedback throughout the whole process. So we didn’t have any one person working full time on TikTok, but we also had more time to set aside for it compared to someone who tries to tackle development by themselves.
So while many of my learnings from this journey should apply to just about anyone setting off on their own TikTok ventures, I ask that you take this background into consideration when you look at our results. (With that said, I see much larger brands fail to get the greatest results on TikTok, so we’re likely doing something right).
Getting Started
Studio vs. Game Account
Our first debate was whether our account should be named after our game or if it should be a more general studio account.
The pros of having a game account are that people are more likely to be searching for our game by name rather than searching for our studio. The brand recognition momentum that we have is currently for the game itself, so we were worried that the username “Inflexion Games” would have less of an impact.
The pros of having a studio account are that you can start to build that studio recognition over a longer period of time. Studio or “day in the life” content may also feel a little more at home compared to a game-specific account.
Given that our current objective was to focus on growing the interest in our game over long-term studio recognition, and to keep it consistent with our existing social channels (which had been established before I joined), we decided to go the game account route. We also decided this as we have the option to change our account name later on if our studio decides to work on a Game 2 (or 3… or 4…) so we’re not trying to migrate an audience to another account later on down the road.
Maybe if we had seen TikTok as less of an experiment and more of a long-term strategy, maybe we would have considered the Studio route more seriously, but that’s something for us to worry about later on down the road.
Account Type
The next consideration was whether we immediately wanted to set up a Business, Personal or Creator account. To see the specifics of the differences in each account type, check out this page here.
We really wanted more detailed analytics to measure our content experimentation, so that ruled out the basic Personal Account.
We knew that choosing a Business Account would limit our ability to jump in on many trending sounds, but as someone who has received monetary penalties for social content before, I was very much on the “once bitten, twice shy” side and suggested we take the more legally safe route.
I know a lot of brands, big and small, who use Creator Accounts, but I will very much leave that to your personal risk tolerance. Sure the chances of being taken to court are low (the sound is more likely to be removed from your video, or you’ll be asked to remove the video), but lawyers are extremely expensive if you do get caught! (And a Business Law 101 class I took in university told me there are additional implications of using trademarked sounds to advertise a business or service compared to individuals just playing around with an app [CITATION NEEDED, a lawyer is your best friend here]. Also, laws vary widely per region… but that class royally spooked me and is one of the very few things I retained from formal education.)
Tone of Voice
The next thing I did was confirm our overall tone of voice for the content we were going to create.
Generally, our marketing content has leaned towards a more serious, mysterious, dark and fantastical side. And I wanted to largely keep this vibe going on our video content.
But anyone who uses TikTok knows that authenticity is incredibly important on the platform. So I tried to leave little hints in captions or write comments that were a little goofier and gave little winks that I was a TikTok user myself.
VO vs No VO [Ads]
For our first couple of videos, we knew that we wanted to do a general game overview to introduce folks to it if they hadn’t seen it already. We planned to pin these posts so that if later on another video caught someone’s eye and they went to our profile, they could quickly get information about our game.
Now to get a little personal, I have little to no confidence in my voice. It’s so far from a nice radio or anime mommy voice, so I wondered if it would harm our video performance. We went back and forth as to whether we should get another team member to record the voiceovers, or if we could get a contract voice actor for them.
After some discussion, we decided those options would greatly affect how quickly we could make videos, could create bottlenecks if we ever wanted to jump on a quick trend before it went stale and of course, in the case of a professional VA, could cost us extra money.
So we decided to make two similar videos, one with my voice-over (VO) and another without VO to see what would perform better. These initial videos were struggling to get over 1,000 views so we put a couple hundred dollars behind each one to expedite our testing process.
The results were clear.
The version with VO, even if it was my own voice, had more than double the video completion rate, and engagement rate, as well as getting higher views for the same amount of money as the video without VO.
So our path was clear: we absolutely should go forward with voiceovers and not just subtitles. And maybe my voice wasn’t so bad after all.
Face Cam???
My immediate hypothesis was that people like seeing other people’s faces and having some kind of green-screen overlay over our game footage would perform better. Or at the very least feel more personal than just pure gaming footage.
But after talking to a handful of other community managers, it turns out they had already done some experimentation here, and they found out it was the opposite! Their videos performed better without someone appearing over game footage and since adding that type of overlay added so much more to production time, it was a no-brainer to remove that step altogether not only for them, but for us as well.
So this is why you won’t see my face in the majority of our videos!
Building a Video Backlog
There’s an urban myth that TikTok gives a visibility boost to new accounts for the first couple of weeks, so folks should post more often during this period.
Now, I talked to a TikTok representative at Gamescom, and they told us that the algorithm doesn’t actually do this. But my suspicion is that the algorithm doesn’t know enough about you or your content so it shows it to a much broader audience compared to later on when maybe they have a lot more data based on who watches your videos and follows you.
So even though it was unverified information, we decided it wouldn’t hurt to at least try to post as often as we could within that first two-week period, and then scale back our production to a more manageable pace for our small team.
We planned to post ~8 videos in the first two weeks and built out a plan to prepare these videos in advance rather than make them in a frenzied panic the day of. After that, we would reduce it to one or two videos a week.
Our plan for the first batch of videos was simple:
- Make a couple of game overview videos to pin to the top of the page (while also test out VO vs. non-VO)
- Create creature overview videos as they were often the focus of YouTube comment discussions while making these videos feel as organic to the platform as possible
Setting Up Viewer Interactions
I figured that TikTok, like many other social platforms, would reward viewer engagement. And it has since been confirmed by the CEO himself.
Inspired by videos like this one by Jared Tan, I would find ways to encourage comments either in the videos themselves or in video captions.
While I didn’t want to overuse this tactic as it can come off as kind of cringe, there are studies out there that say people are more likely to do something if you directly ask them for it.
I decided to kill two birds with one stone:
- Find out which creature was the most interesting to the TikTok audience
- Directly ask them to comment on the video
So I made this video (yes, there’s no VO in this one, it was filmed, edited and released before our tests had concluded).
While it doesn’t show all of the creatures in our game, I made sure to include the creatures that I personally thought were our game highlights and also scraped through our trailer comments to see what people were mentioning the most. Then I added a clear call to action in the captions.
Admittedly we didn’t actually wait for comments to come in, we were preparing all of the creature videos in advance starting with the one I thought would be the most popular. I got it mostly right, but we did bump some of the recordings earlier to accommodate the organic traction we were getting with this video.
From here I used the TikTok feature that lets you reply to comments with our prepared creature videos so that user comments would show at the beginning of the video. This would then encourage people to comment on our original video in hopes that they could learn more about the other creatures!
The first video reply we posted was the Carnute and it proceeded to get over 1M views within 24 hours!
Comment Interactions
Something I did for every video, was sit and respond to comments for the first 30 minutes after something was posted.
I don’t know whether TikTok counts the video owner comments towards the algorithm, but something I noticed was that people were more likely to comment if they saw that we were responding to folks!
I wouldn’t respond to everything, in fact, I would ignore quite a few comments (especially ones that compared us to other games or generally exhibited behaviours we didn’t want to encourage), but I would make sure to give thanks to people who expressed their excitement or answer as many questions that had publicly available information as possible.
But beware! If you have a viral video (getting hundreds of comments in the first couple of hours), you can get flagged as a bot and locked out of your account for a day due to responding to too many comments. I managed to do this even when I was hand typing out “Thank you”s and little hearts to folks, even with breaks to work and eat!
Video Looping
Early on we were inspired by perfectly looping videos like this one.
We saw a correlation between video completion percentages and performance, so we tried to hide the fact a video was ending by splitting our video clips from the end so they continued to the beginning video like in this video.
While we don’t always use this tactic, we try to do it as much as possible.
Analyzing Our Content Successes
Our best-performing videos early on were the Carnute and Harpy videos, currently sitting at 2.4M and 3.6M views respectively.
So why do I think these ones did particularly well?
Well, I think it’s for a few reasons.
One: The designs are eye-catching
I’m going to toot our art team’s horns here. From the concepts to the in-game models and the animations, I think our creatures are quite cool! And while some of them might draw comparisons to other media depictions of creatures, I still think ours have their own original flair.
Our creature models are intricate, detailed, and definitely at a scale which fits our game, studio and engine (UE5), but this might be hard to apply to your own use case depending on the type of game you’re making.
Two: Appeals to a wider audience
Something I was targeting with these videos was making them so they didn’t just appeal to the gaming crowd. I knew creepypasta has been popular in online spaces for years (decades?), so I leaned into those aspects of our lore.
I also don’t mention our game at all in these TikToks! I made a conscious decision to just talk about the creatures and lore to make them feel more organic and less sales-y.
For those who were interested and clued in that it was game footage, they’d ask was game it was (even though it’s in our name…. our bio…. and in the hashtags….).
Of note, our creature videos generally perform above average compared to the rest of our content, so they’re definitely resonating with our audience!
Three: Comparisons to other media
It wasn’t intended, a lot of people commented on our Carnute video that it reminded them of the historically inaccurate, but popular media depiction of a Wendigo. At no point in the video do we talk about it, and when people asked what inspired it, we openly shared it was lightly inspired by, but not meant to be a depiction of, the Celtic god Cernunnos.
While we didn’t purposely encourage such comparisons, it is very likely that we benefited from these comments boosting our spot in the algorithm, or at least inspiring people to watch the video due to perceived familiarity.
I don’t recommend purposely trying to co-opt problematic media depictions of things, but it’s hard to ignore its impact on this particular video.
Then for the Harpy video, a number of people mentioned how our Harpies don’t match up with various other media depictions of harpies. We knew that we took some creative liberties with our creatures, but canonically Nightingale is an alternate-history game, and our creatures aren’t meant to be accurate depictions of myths, but instead maybe a creature a human saw that inspired the actual myths behind them. Early history is full of creative interpretations of creatures after all.
So again, while not this video wasn’t intended to get “explain-bait” comments, comparisons to other media absolutely helped the Harpy video.
Hashtags
There was some debate in the community space as to whether hashtags helped or hurt TikTok video performance. Some felt like it pigeon-holed your content to certain audiences, others thought the algorithm saw it as spammy and would punish you for it.
I looked at a bunch of viral content, both in the gaming sphere and ones outside of it. There wasn’t a clear consensus as to how many hashtags should be used as they were all over the place, so I opted for 5 or less. Enough to try steer the algorithm towards people who were more likely to watch and engage with our content, but not enough to look like bot spam.
Generally, I would say if you’re going to use hashtags, be picky with which ones you select. The algorithm rewards you for completed video views and engagement, so you want it to be put in front of people who will watch it. If you put it in front of irrelevant audiences, you’ll likely get punished for it.
I’d recommend trying to pick 10 hashtags or less, and try to keep them as relevant as possible rather than too generic (like #fyp or #trends).
What Hasn’t Worked
End Slates
When we first started our TikTok journey, our overall marketing focus was on securing Steam wishlists. With our first creature videos, there were concerns that users were watching our creature videos, but didn’t realize it was from a game.
Utilizing the same psychology principles that I mentioned before that people are more likely to do something if you ask them to do it, a request was made to add an endslate to a video or two to see if it encourages people to wishlist.
The result was staggering.
There was a massive bounce rate before the video ended, right when the endslate showed up. The video was basically dead in the water and barely got 2k views within 24 hours while our other videos were easily getting a minimum of 10k-1m views in the same timeframe.
In the spirit of experimentation, we removed the endslate and reposted the exact same video. Lo and behold, the exact same video would go on to be pretty viral (500k views in <7 hours) and continues to be our best-performing video of all time!
Multi-Part Videos
We had a lot to say about Realm Cards and wanted folks to stick around to learn more.
But the second we mentioned a “part 2”, viewers bounced from the video greatly hurting our video completion rate and what felt like our video performance as it performed significantly below average at the time.
I’m a TikTok user myself, so I should have known the whole “normalize putting all the information in one video” meta, but at the time we were worried that users would bounce from longer form (>30-second) videos so we thought we’d give it a try.
Lesson learned! It’s unlikely we’ll ever cover every detail in a single video, but we’re going to utilize playlists and avoid mentioning future parts or videos in TikTok itself.
Results
So maybe you’re sitting there like “okay that’s quite a lot of views… but so what?”
Well, our game isn’t out yet so we can’t talk about how views translate to sales, but we can say that on videos which performed well (100k views or more), we could see our daily wishlists increase by anywhere from 10-40% higher than average. Not bad for videos with no ad spend and probably <5 hours of work into them. And these were during periods of time where we could directly attribute this rise to TikTok because we didn’t have any other activities happening in those time periods.
It’s worth noting that other developers have seen measurable business impact thanks to the platform, so it’s well worth exploring (assuming it doesn’t get banned in the near future).
Take Aways
Okay, that’s a lot of words. So what are the TL:DRs that I think you should apply to your own TikTok journey:
- Focus on the content: There’s not one singular way to succeed on the platform. This video doesn’t have audio and still got 1.8M views. We don’t use trending sounds and still managed to get views with 3.2M views! Don’t feel like you’re pigeonholed into making any one type of content.
- Be organic: TikTok users are savvy, and they’ll swipe if you come off as too sales-y. Make content which is interesting and authentic to the platform while finding ways to let your product talk for you.
- Hook: Make sure that the first video frames and dialogue lines (if applicable) are the strongest in the video. You have <2 seconds to grab someone’s attention and keep them watching.
- Bite-sized content: Don’t make your videos too complicated. Pick a single topic, and use as little time as possible to get your point across while still making it entertaining.
- Have purpose: Focus on what makes your game (or studio) cool or unique. Don’t post content just for the sake of posting content if it’s going to be boring, dry, or completely irrelevant.
- Continuity: Try not to indicate that your video is ending to the users. Avoid long pauses at the end, endslates with calls to action, and don’t mention “part 2”. Instead, try to create seamless transitions to the beginning of the video wherever possible.
- Encourage Interactions: Create content which organically encourages interactions, and don’t be afraid to occasionally blatantly ask for it in the videos or in the captions. Asking questions or for input are great ways to do this!
- Engage: Spend some time responding to comments as folks may feel encouraged to comment themselves if they think you’ll answer them.
- Experiment: Don’t be afraid to try things! Sometimes things work, sometimes they don’t. The great thing about social media is that we’re all learning together, so take some time to re-evaluate your content, make guesses as to why things did or didn’t work, take notes of content both in the industry and outside of the industry that’s performing well, and keep making content.
- Don’t Worry About Posting Every Day: If you’re short on time and resources, focus on the quality of your content rather than pumping out constant content. There’s a balance to be found where you want the content not to be too sloppy so it’s good quality, but also don’t spend too much time on it so that you’ll be majorly impacted if it flops. We post once or twice a week and have seen really great results! This account has posted 17 videos since August 2021, and they have 1.1M followers!
What’s Next???
We’re still early on in our TikTok journey, and there’s a lot that we’ve yet to explore.
- While we haven’t yet done trending sounds, we want to try them when the time is right (and available to our Business Account)!
- TikTok has a built-in Q&A feature which we’d like to try out and see how it goes, but we’ve been hesitant as each answer requires a video, and we’re focusing on other content at this time.
- We haven’t tried out filters, either using existing ones for our own video or making our own filter for others to use. But they seem like they could be fun.
- We’ve talked about showing off some of our more personable sides with more IRL content like showing off the dogs at the studio, or maybe game dev interviews… but these two have taken a backseat to prioritize in-game content at this time.
Hopefully, I can report back in the future with the results of such content!
Shout Outs
I would not have been able to make this journey without others who graciously shared their own experiences on the platform. A huge thank you to:
Jared Tan
- Cult of the Lamb TikTok
- TikTok Optimization Twitter Thread
- TikTok Strategy
Victoria Tran
- Among Us TikTok
- Unpacking TikTok Strategy
- Among Us’ TikTok Stategy
Hanna
- Landfall Games TikTok
- TikTok Twitter Thread
- Update to Above Thread
Thomas Reisenegger
- TikTok Guide for Indie Devs
Non-Game Dev Accounts I Think Are Neat
DuoLingo
Bose
Ryan Renolds
MadMorph (Game and movie summary account, 930k followers!!!)
