- ARPDAUPosted 5 years ago
- What’s an impressive conversion rate? And other stats updatesPosted 5 years ago
- Your quick guide to metricsPosted 5 years ago
[Gamesbriefers] Low to high, or high to low?
A GAMESbrief reader asked whether he should list his virtual currency from most expensive to least expensive in his store, or the other way round. What advice do you have for developers thinking about how to lay out their store?
I think order itself is less a concern than the whole presentation: Because an IAP is a non-generic good the perception of value is created through comparison, so how you display that comparison (and what the comparison is) is the important bit. In my book I go in to a bit of depth on this and use Dan Ariely’s Economist subscription experiment as an example (see his TED Talk on it here).
So, I think you’re right on highlighting “most popular” and “best deal” as is commonplace. However, I don’t think a linear increase of savings is necessarily the best thing – having a “sacrificial” IAP which is poor value comparatively can drive purchases higher.
However, if you can multivariate test the story, do so – until you have that data you’re just guessing. If this is an iOS title you can’t vary IAP prices easily (or at least without creating inaccurate data tinged by player comparison to old prices). You might set this up so that the IAP price doesn’t change, but package size does, from this you can do some regression analysis and work out the optimal packages.
One of the striking things about running an F2P game is how much it looks like retailing. I would value the advice of someone who runs a high street store. When should you do a two-for-one deal? How long for? If you cut this price, should you raise that price? How, indeed, do you position your offers.
I don’t think there’s a single answer to this, just things to try. AB test different combos. Engineer the screen so the ordering can be based on player stats – e.g. if you have someone who seems happy to pay, that suggests a different layout than a long-term non-payer, who might be more swayed by a very cheap intro offer. Think about offers presented in the shop (like currency packs) versus in-situ, one-off IAPs, if you have them. I like Nicholas’s advice to present the player with a compelling early purchase, like the coin doubler, to get them to pay something. You might present offers differently for those people.
First: do not offer too many price points. Users get confused. 5 is a-ok, 7 max.
Second: remember that for some formats you have price points where you cant change the price (SMS)
Third: make one point attractive enough so its a no brainer to buy; that will be your ARPPU (see Dan Ariely)
Having said that, we should separate the in game item shop and a shop to sell in game currency. In the item shop, you won’t sort items by price, never, you sort them by category or game specific attributes.
As for most expensive first: I disagree. If a user who is on the edge of paying sees a $99 price point he will likely not pay simply because just the existence of such a price point will turn him away. Only engaged users will use that $99 anyway, so why display it first.
Other responses here are far more scientific, but I remember having this discussion with someone a couple of years ago before we launched our first F2P game, and the two bits of advice were:
a. Order items from largest / most-expensive first.
b. Whatever your largest IAP is, add another above it.
For (a) the logic was simply because people read from the top, so by the time they get to the cheaper items they’ve at least had a look at the more expensive packs to make comparisons.
(b) was to make the mid level items look more appealing – like the restaurant wine list psychology – people tend to go for something ‘mid-priced’ whatever that may be. Looking back, (b) can be seen as a bit cynical. We did add a level above, not realistically expecting people to buy it, but of course people did.
The points made by all about Anchoring are very useful.
It is worth pointing out that everything should be purchased by in-game currency. Even if it’s a one-off, permanent upgrade, ability or object, price it in the IGC and make sure it leaves some IGC left over from any of your IGC packages, to allow people to sample spending, alongside the ‘no-brainer’ purchase.
I feel obliged to add –
I think this is a bit of a dangerous question.
There are general guidelines of course, but this stuff is *super* important and I’d be wary of offering blanket solutions. Teut’s point about $99 being the first number people see might be right. But $5 is probably too low. So where is the exact point? That will depend on a *huge* number of issues, and change from individual to individual, (a fan who buys at $49 might be up-sold to $99 with some smart design) and getting it right will have a significant effect on revenue.
More generally, always think about all revenue related questions early and often and forever and test and iterate and aim to offer great value, not just raise revenue. When you make changes to your revenue model, make bloody sure you’re tracking what it does to acquisition and retention. Selling lots of an item that makes people stop playing your game, or recommending it to their friends, isn’t good for business.
Aside: I’d love to see the numbers on users buying ‘boosts’ for resource collectors in Clash of Clans. They are definitely less efficient than just buying currency directly, which is also possible, and so would seem to add confusion and offer lower value. Which doesn’t seem sensible. But I’m loathe to accuse Supercell of getting things wrong 🙂
That’s a really good point about making sure everything is purchased with in-game currency. It also allows for meaningful integration of incentivised video ads, and other revenue tools. Personally, I really don’t like things like ‘Sign up to NetFlix for 3000 credits’ offers, but incentivised video ads are a pretty seamless solution where everyone’s a winner.
Well, CSR offers one-off large IAPs in the fiction for cash, not in-game currency, and they work very well. Requiring that purchase to be in Gold rather than cash means friction for players who don’t have it – if they want to spend a couple of bucks they would rather do it now, not in two steps.
Do think about the granularity of the virtual currency, though, to give you most flexibility with smaller-scale purchases. But that’s a different conversation.
I think this is exactly the kind of debate which gets us into trouble.
The focus is just on how to make as much money as quickly as possible – and if that is your objective then of course Charlie is spot on with his strategy. Actually his advice is also largely right from a retail strategy point of view too. But from my experience it isn’t the right answer if your plan is to build longer-term lifetime value.
I feel like I’m about to tell that old joke about when you ask for directions only to be told “Well I wouldn’t start from here”
To me the frame of the question is wrong.
What value are you selling to the player? What is the lifestage of that player? Are they capable of appreciating the benefits of volume discounting or will they instead see your game as a money pit?
I think Will has a raised a good point about the comparison when assessing the value of virtual goods but part of this I would argue is about the way we communicate or foreshadow the potential value of that item.
If you show me a list from most expensive to cheapest before I care about the game I will churn. If you don’t show me discounting options when I am ready to spend as much as possible I won’t spend as much on your game as I might have done.
There are so many opportunities to use effective retail techniques to do more than just generate more short term cash. For example when I ran the 3UK platform we found one-time sales to be completely useless. We saw a 3x increase in revenue for one month followed by a sharp drop in income for the next 2 months. In the end we saw that the average revenue per user over time remained unchanged. But when we instead built a process where every game launched would be reduced in price 6 weeks after it was released (from £6 to £3, then later to £1). More than that we decided to remove the game from the deck a week later after it was reduced to £1. These actions set up an expectation of value which made sense to players at all life stages. Through these techniques we saw players making the decision to increase their average monthly spend.
So what’s is my point? Lets work out how we add value to users and find the best way to communicate that – not just rely on lazy lists where we can’t decide on the best order. We are all retailers now.
The only answer is of course ‘run an A/B test’.
“One accurate measurement is worth a thousand expert opinions”
I agree completely with Ben, and will throw in a related quote:
“If we have data, let’s go with that. If all we have are opinions, let’s go with mine.”
Two quotes come to mind for me too…
“Its not about the money…”
…well not just the money – and in Ben’s example their use of an effectively premium currency does infer value to the long term life of the game (I just didn’t think it was enough value for me)
“Lies, damn lies and statistics”
…meaning we can over rely on the data we can measure, as well as misinterpret its significance. And of course 1000 expert opinions collected properly could constitute data :0)
Ben Cousins Head of European Game Studios at DeNA
“Should our virtual currency offers be ordered low to high or high to low?” seems to me like exactly the type of thing which can be tested in a data-driven way with very little chance of noise.
‘What features in our game make it the most fun?” would be an example of something that you maybe wouldn’t get great answers from using data analysis.
My comment just referred to the original question. IMO there’s little point in developers wasting energy trying to figure out what to do, just test it.
I think there’s value in questioning what’s worth testing, because it takes time to create a multivariate tests and blindly running them is benefit to nobody.
This strikes me as exactly the sort of problem that A/B testing was meant to solve. Your reader needs to focus on is building his store in such a way as to be configurable server-side (and therefore avoiding an App Store approval every time he wants to make a change), A/B-able, and then try a few configurations to see what works best. As no doubt everyone will say, there’s no ultimately right answer, so whatever works.
I don’t think anyone is suggesting that a/b/n testing is the only answer, but would any of you not use an a/b test to answer this question?
‘should we list our virtual currency from most expensive to least expensive in the store, or the other way round?’
Why wouldn’t you? Because it might not be the right question and waste of time.
ABTs are great and necessary, but not the only answer. In the universe of possible options you have to start somewhere. Every ABT you do takes time (in development and in waiting for results), money and opportunity cost. You have to apply some smarts to choose where to start.
I agree. I’m obviously looking at the question isolated from any meta-questions.
In an effort not to reiterate the many good observations already made, I’ll keep my reply to two specific examples.
Pixel People has an IAP that doubles the timers of all buildings. I would have purchased this fairly early on, after perhaps 4 or 5 game sessions at which point I could really see the value of such an upgrade. However the upgrade was only purchasable within a specific building, and entirely by chance that building was literally the very last building I built (out of the initial set), at which point I had no use for it. While perhaps “cute” to design purchasing within a building, the reality is that it is a poor strategy financially and it is a prime example of the points made earlier about surfacing and targeting players appropriately with offers that make sense with where they are in the game lifecycle.
At Kongregate we do see players balk, sometimes very vocally, at seeing high price points in games. When you’re considering putting your first $10 or $20 into a game, seeing a $500 price point is not only intimidating but also minimizes the perceived value of your first purchase. On our site, and in some of the games we work with, we instead choose to suppress large price points (generally above $100) from players who haven’t purchased before. After an initial purchase, especially above a certain size, we can then reveal that larger price point because it now makes more sense for the player. While it doesn’t solve the bid-endian / little-endian debate, it does provide an additional tweaking point that you can work with.