The content material of this text was first introduced by Simon Thillay, Head of App Retailer Optimization (ASO) at AppTweak, through the ASO Convention 2021 on June third. This spring, the App Monitoring Transparency (ATT) framework was launched as a part of Apple’s efforts to advance consumer privateness on the App Retailer.
We’re all happy to assume that in order to grow a successful mobile game (or app) you need:
ATT now requires app builders to acquire specific permission (opt-in) from customers to trace their information or share their info with advertisers. Consequently, successfully focusing on the appropriate audiences with cell adverts on the App Retailer has change into extra difficult. As a substitute, many app builders are contemplating shifting extra promoting budgets to Apple’s native advert platform, Apple Search Adverts (ASA), as a consequence of its exemption from the ATT limitations. On this weblog, we offer a mannequin that can assist you to measure the price of cannibalization in your Search Adverts campaigns and perceive the professionals and cons of Apple Search Adverts.
Apple Search Adverts in 2021
On paper, the ideas behind Apple Search Adverts are fairly easy: app builders can bid on any key phrase to have their app seem above the primary natural search consequence for that key phrase, in hopes that this outstanding place within the search outcomes will assist to extend app downloads. The App Retailer algorithm supposedly determines which app wins a bid for a key phrase by contemplating the metadata (title, subtitle, screenshots, and many others) and efficiency of every bidding app.
Nonetheless, the chance for any app to bid on any key phrase – together with model names – has led many apps to bid on their opponents’ branded phrases. In flip, this has pushed apps to bid on their very own branded phrases to forestall opponents from ‘stealing’ their visitors. The next query due to this fact arises: in defending your installs and impressions from being ‘stolen’ by your opponents, what number of of those installs and impressions would your app have gained organically? That is visitors that has fallen sufferer to ‘cannibalization’: once you pay for visitors or app installs that will have in any other case been pushed organically.
Modeling the price of cannibalization vs model safety
To mannequin the worth of cannibalization, it is best to first perceive how we outline Cannibalized Installs. Cannibalized Installs are the whole variety of installs you acquired out of your Search Adverts campaigns minus the variety of Protected Installs.
Cannibalized Installs = Complete Installs – Protected Installs
Protected Installs are the variety of Installs that you just acquired, however that your competitor may have been in a position to convert into downloads.
Protected Installs = Impressions * Estimated Competitor Conversion Price
The Price of your Protected Installs then turns into:
Price Per Protected Set up = Complete Advert Spend / (Impressions * Estimated Competitor Conversion
Price)
Evaluating your Price Per Protected Set up with the Price Per Set up reported by Apple Search Adverts can then assist you to estimate the ‘safety premium’ you pay for a given marketing campaign; this may even assist you to decide what quantity of your Search Adverts spend truly goes towards cannibalization. Sadly, figuring out these prices doesn’t take away the dilemma of whether or not it is best to bid in your prime natural key phrases or not, however it could possibly at the very least assist inform your choices to begin, cease, or change the finances you put money into particular Search Adverts campaigns.
Precisely estimating opponents’ conversion fee
To find out your Price Per Protected Set up utilizing the above equation, you will need to perceive whether or not the competitor who ‘steals’ your paid impression would be capable of convert it into an app obtain, thus estimating your competitor’s conversion fee.
You need to first formulate conversion fee hypotheses. That is the ‘hit and miss’ a part of the complete cannibalization mannequin: on condition that conversion charges are intently guarded information (Apple doesn’t even share with builders natural conversion fee information at a key phrase stage), it’s a tough problem to formulate correct hypotheses about conversion charges. As such, the perfect route might be to make use of proxy variables – substitute variables that assist us approximate conversion fee – to make sure your estimates aren’t utterly unrealistic.
The principle information factors to make use of in such a case are usually the keyword-level conversion charges that you would be able to get hold of from your individual Search Adverts campaigns. It is just logical to think about that your app’s conversion fee to your personal model identify is unlikely to be decrease than the conversion fee of your opponents which can be additionally bidding in your model identify. Subsequently, you should utilize this determine as an ‘higher bracket’ that opponents may have issue reaching. Then, you may invert this logic by contemplating your individual app’s conversion fee for a competitor’s model identify; this can be a nice technique to anchor your speculation and higher perceive how your competitor would possibly convert customers that seek for your model identify.
Upon getting decided your first anchors, it is best to calculate a similarity rating between you and your competitor(s), and refine your conversion differential hypotheses based mostly in your consumer analysis and information factors introduced by third-party instruments. Specifically, keep in mind that the app shops are strongly grounded on relevance – the a number of locations through which your app seems on the shops characterize simply as many alternatives to match your similarities and conversion performances with different apps. For example, the highest natural search rankings are inclined to mirror which of essentially the most downloaded apps convert higher than others for particular key phrases or queries; these search rankings are due to this fact nice hints that will help you estimate the various conversion charges of your opponents.
Concerns when estimating conversion charges
There are a selection of various elements you will need to think about when estimating your opponents’ conversion charges:
- It’s firstly vital to contemplate an app’s model energy and client loyalty: when trying to find a sure question (branded or generic) on the shop, totally different client segments have totally different ‘first selection’ apps in thoughts. Assembling a consumer analysis crew might help you higher perceive model notoriety on the App Retailer and offer you an preliminary thought of your app’s model energy; nonetheless, don’t neglect to make use of each quantitative and qualitative analysis strategies to measure how customers react to your model. Manufacturers that prospects establish strongly with (as an illustration, due to a way of shared values) are lots much less prone to have prospects change to a competitor’s product when given the chance.
- Subsequent, it is best to think about your opponents’ notoriety: opponents that almost all prospects have by no means heard of are inclined to characterize a smaller risk to your model visitors. Nonetheless, even when your prospects are unfamiliar with a competing product, you will need to additionally think about the perceived similarity of this product to your individual. Your merchandise may very well be very totally different; nonetheless, these perceived similarities characterize the best danger of retailer customers adopting opportunistic behaviors and reconsidering which app to obtain.
- Third, it is very important think about shoppers’ true search intent. For example, shoppers trying to find “[brand name] free” might merely be trying to find an app that provides functionalities just like the branded app however for a less expensive value. On the opposite finish of the spectrum, sure manufacturers with extraordinarily excessive notoriety and a robust presence might as an alternative characterize a complete style of apps. For instance, the model names “scrabble” and “mario kart” have been regularly looked for on the App Retailer even earlier than their official apps have been launched; this demonstrates that customers don’t at all times seek for model names with the intent of downloading that app, however usually with the overall aim of discovering the same product.
Use the cannibalization mannequin effectively and cautiously
Measuring the precise price of cannibalization has been central to Apple Search Adverts since they have been launched on the App Retailer. Nonetheless, it is very important keep in mind that fashions, such because the one introduced on this article, are designed to simplify actuality as a way to facilitate decision-making. When utilizing this mannequin, don’t neglect the next issues:
- This mannequin doesn’t account for sure parameters, akin to the share of customers who mechanically ignore apps in Search Adverts placements, or the truth that not all customers obtain an app after each search question on the shop.
- The mannequin’s accuracy will rely upon the precision of your hypotheses. Consequently, it is best to often take a look at and retest Search Adverts campaigns to maintain the info behind your hypotheses updated. As with natural performances, Apple Search Adverts campaigns are carried out in a continuously altering surroundings, and each key phrase or artistic change – in addition to many exterior elements – can change an app’s conversion fee day by day.
Extra app promotion guides: Make the most out of creative assets to optimize paid UA campaigns in iOS 14