Space Ape on the power of CRM for F2P success

so yeah thanks Mitch for for that overview I know your GDC talk inspired a lot of us at space ape to to rethink this and on behalf of everyone the on the brutal buy side of this equation any efforts to increasing battery as is great by our so thanks so topic of zero of this month is my clock okay it's a topic of of my talk with Billy and analytics tak a little bit about us space ape I'm the CEO and Founder based in London we're about a hundred people studio three years old split across four game teams to live samurai siege which is just celebrated two year birthday and rival kingdoms which is about four months old both both in the combat

city-builder genre and yeah we very we very much depend on live ops and content releases to give these games longevity so we'll be putting new events new content every week and so the games really really depend on a strong analytics infrastructure so yeah so when George pitched me about coming in and giving this talk is really referencing a talk that we gave in London about five or six months ago where we we open the kimono on our analytics stack and when I was trying to figure out how to make a 15 minute version of that it got a little bit tricky and so I thought I'd talk about one or two topics within it within that presentation but if anyone's interested in getting the full the full deck just Google space ape analytics stack it's the first three results up it's about a 60 page PowerPoint presentation and I go through all the details of how we arrived at their current stack but in there you'll see

beautiful graphs such as this one would show the architecture and really the the point of that talk was to show how we evolved as a from a 12-person startup who really couldn't afford to put too much investment into analytics at all yet knew that it was important through to the hundred hundred dollar person company and really when we started we just had a basic swerve implementation in samurai siege in beta but an as we grew we kind of evolved it so at the start we made sure that we still had all the raw events dumped in an s3 bucket somewhere so we knew it was time to hire an analyst when myself and the product managers are spending more time scraping those logs - to answer questions and then we we hired a data engineer who started building the air to your processes and outsource a bunch of stuff and eventually grew it into into a larger stack but the the main point is we only built what we needed at the time and and but with it with an eye to -

what we need in the future in that deck you also see a breakdown of our single player views so if you're if you're trying to write up all the event taxonomy for your game right now trying to figure out what you should track this is a good reference - you could probably reverse engineer our event structures tags from this from this deck so encourage you to have a look key takeaways were something that we didn't see much two years ago when we did this make sure you keep ownership of your or event logs so if you're using a vendor make sure you have access to it so you you can then build on top of that and only build what you need and make sure that that raw data is accessible throughout the whole company so a lot of the innovations and the shortcuts that we took on data we're a result of everyone having access to it and and riffing on that so for example our CRM system their player profile system was a result of one of the server devs just seeing that this information was there and trying to make it more easy to interface with one of the topics that

came out of that presentation that I get a lot of follow-up questions on was one of the the ways of this data enabled us to do some pretty interesting marketing techniques this is a screenshot from the very first beta of samurai siege our first game and it was live in New Zealand it's pretty embarrassing UI right now but we put it live early to get feedback and on this screenshot you can see a hint of probably our most effective marketing channel that we we stumble across more effective than the UA arbitrage that we invest a lot of money and more effective than the featuring that we got great as it was and that's that little blue button in the corner and it wasn't obvious at the time but the state of customer support in Mobile was was atrocious at the time it was impossible to get a response from any developer really for good reason and and what we did is we didn't have any grant strategy on this but we wanted to test a bunch of features we're

introducing Alliance Wars and some different event formats and we really didn't know whether it's going to work we want a cheap way of getting feedback so we we worked this button up there and I was responding to the emails and in the first week it became very obvious that we're on to something I would reply to someone who's at I you should do X Y and Zed nod reply and then I get a response back almost immediately again oh my god I can't believe the developer replied I didn't even say was but I said I was the founder of the company it would have would have blown their mind but but the fact that anyone in the company and responded to them was was mind blowing and so a week of so went by and then we started diving into all of our top players and the Alliance's that are on the leader boards and we saw a very clear pattern with people who had contacted us had a good experience were blown away by the fact that developer really cared about players said and then went and recruited all their friends and so we saw a very strong correlation between positive engagements through this feedback button and ultimately

spend in the game so much so that in the first six months of of the games lifetime that 20% of our revenue was attributable to people who had had good experiences with customer support now that could be this is a pretty loose data there's a there's a lot of causation versus correlation questions over this but we believed that we were on to something and we really invested a lot in support and community and and it's something we continue to do today some of the ways this manifested were people who were big spenders in other games or influential in other games came along and wanted to to support us as a developer who who listened otherwise it would manifest would be people who were on the verge of of giving up over an issue obviously the classic customer support resolution type problem but but all tied back to this feedback button so going back to the the the one of the users we put to the data was we made

sure that every CS ticket that came in was augmented with all of this CRM profiling and propensity data so when someone wrote in to us from the game any agent who picked it up can see obviously the usual metrics around spend level activity engagement but also stuff that the user couldn't see such as crash logs that they may not have known were being stored and which helped the agents diagnose the problem but also their propensity to churn their propensity to spend where they were in their lifecycle whether they associated with other valuable players and that really helped us triage the tickets because once we got to about 10,000 daily users you know beta it was impossible to read our CS couldn't couldn't handle that with with the grand headcount of 1 so we we had to triage in and so what we found is we were funneling people through to have a good experience that the data was telling us would be more valuable and and so that was one of the ways that

once we were able to hook up all of the systems without with our CS agent tools it became really powerful another tool at our disposal was there was their in-house customer profiling section so so when a ticket came in you could easily open up and see all of the interactions that that player had had initially it didn't look anything like this there's more like four fields but what we did continuing the spirit of making sure those roar event logs are available to everyone we were just one of those fields was adjacent dump of everything about the the player's profile one of them was a log stash link so more technical people could could could investigate things and then over time we retooled it up a bit better this is my profile in in samurai siege so one thing we we we said then is look how do we scale this is clearly something that was working for us at a

very small scale with a very small number of users in beta how can we blow this up I had four or five people we just said okay we'll try and make a community fallen out of it turns out this was a complete waste of our time it was very valuable for the small number of people that engage with forums but what we learned was actually our most valuable and most viral and our most influential players we're not going to go to forums it's a pretty strange experience if you're not a foreign player they also will not really engage with their friends and you couldn't really track this reality but what they were all using was chat apps to keep in touch with each other this is strapping outside the realms of the data but as one of the things that the the data informed us on to lead us down this path when we when we started in this genre the top thousand or so spenders in in the genre they work very viral they'll keep in touch with their friends but there were friends they met in the game and about 13% of them were using line 11 percent of them were using

WeChat and kick and and other other chat apps are up there as well fast 40 year nearly 30 or 40 percent of these people were online and so that led us to scrap the forums and create live line chat groups where if you if you qualify as a player according to our data you you get into these you get invited to these rooms where you've got basically 24/7 live chat support and that is our community so rather than have forums all of our community efforts are geared around these chat rooms in applications that our players are all residing in so we call it a VIP program staffed by country managers they run localized events they give sneak peaks and access to to information and when a player is in that category of we that we we have marked as as needing will be warranting more investment then that status is visible to all agents we have different tiers and and and that's visible

throughout every time they contact us another data point as a result of this program these are some stats from our beta and we were testing rival kingdoms in March you see there we've got about 5,000 DEA you during this period and we're live in four countries when you actually add up those numbers it come to about 4,000 do you the the other 20% of people playing in our beta were changing their countries too to get access to the beta and that gets even that gets even bigger when you look at revenue so a lot of those people were calling up iTunes support changing their payment methods finding people with gift cards and actually the majority of our revenue was coming from people outside the beta countries and when we approach these people ask them what they were doing and why they were doing it the answer in almost every case was we're supporting space ape as a developed because they listen to us they really value our players and feedback and so tying that all back to the the

initiative has started off in our beta it was one of the the main reasons people going through jumping jumping through a lot of Hoops to get access to our game in the beta so that was it please oh yeah old-fashioned customer service works and main takeaway we do see the effect in virality particularly amongst high value players it does get harder as you scale so at the size right now it's huh it was pretty easy to know that we should put two or three people on this knowing where you should put 5 10 20 or more people on it that's a much harder question to answer and and yeah that's that's it really I'm out of time I think I'm out of time it's pretty hard to read this might have two minutes left or it might be one minute over I'm not sure one minute I've got a minute I got a minute where the questions if anyone wants to ask far away we've been through a few different evolutions of this initially it was just an email that we pre-populated with session information

the client could gain in particular the user ID we then evolved that to some very broad I've got a problem I want to make a suggestion I can't remember the third category is in rival kingdoms we have a pretty sophisticated system now where we we have six or seven categories which are filtered based on your your life stage but yeah we really try and funnel technical questions here feedback and suggestions we try to send people through to a forum with a drop-down list that kind of thing but it's something we're optimizing all the time but yeah the classic examples I've got a suggestion I want some help I can't that kind of thing but the real valuable ones that people have suggestions time for one more question I guess not yep hi how do you do the attribution from from so like these social chat apps to an actual so like revenue for accent yeah

so we can't do that what we do is we do an attribution of people we invite to join so all of our community rooms are invite only it is technically possible for players to invite players but we make sure we always identify who they are so it's a manual tagging process we've got about I don't know what the last chaos between four and five thousand people in this program and we have them tagged in the backend across both games