Under the Radar

Under the Radar 109: Search Ads in Practice


  welcome to under the radar a show but

  independent iOS app development I'm

  Marco Arment and I'm David Smith under

  the radar is never longer within 30

  minutes so let's get started so Apple

  launched their search ads program last

  fall in the App Store and a lot of

  developers I think us included didn't

  really know what to make of it at first

  it we were it certainly was a little bit

  intimidating a little bit off-putting

  the idea that we'd have to pay to have

  our ads shown at the top of the search

  results in the App Store and we've now

  had a lot of time I've actually been

  using search ads for overcast since the

  day they launched and I'm still using

  them today it's a little bit back in the

  news now because yesterday Apple

  launched I think by surprise for most of

  us Apple launched a new search ads basic

  which is a version of search ads that is

  so far us only but they said it's going

  to expand to the rest of the world and

  that's where that's where the ads show

  up not who can buy them so if you're a

  developer outside the US you can buy

  these ads but they will they will only

  start showing in the US App Store for at

  first but anyway so now they have search

  ads basic and the old one has renamed

  search ads advanced I think it's worth

  discussing now that we had you know at

  least from from my perspective and and

  from yours David to you know to the

  extent that you've used it now that we

  have almost I have over a year now of

  search ad data it has slowly expanded

  from initially the main offer it was

  also us only and then it expanded over

  the last maybe six months or so to a few

  to other different marketplaces now you

  can have Canada UK Australia Switzerland

  and New Zealand and it's a little bit

  cumbersome to manage them all honestly

  but I think I've come up with a number

  of key insights and and best practices

  that might be useful to share what's

  your experience with them because you I

  mean you have way more apps than I do

  because like all I can say is what what

  my experience has been with my single

  app and it's a free app so it's gonna be

  different than paid up front apps and

  its affiliate it's you know fairly

  popular among certain people so that's

  gonna be different from an app that's

  totally unknown have you had have you

  had more diverse experience than I have

  with this a little bit so like I've

  tried tried well I've tried a variety of

  different campaigns

  and things with this I tried it for my

  paid apps for my free apps and overall

  like my experience is like in aggregate

  is just very mixed I would say like some

  of the time it it seems to work and pay

  off sometimes it really doesn't like and

  the thing that I struggle with the most

  with these is trying to it's like

  there's initially like the obvious naive

  like reaction to something like


  he's viewing it as a in a weird way

  almost like a direct investment where

  like I'm I take this amount of money I

  put it into this and then I expect to

  get a certain amount hopefully a greater

  amount back and that was straightforward

  especially for like my paid apps for

  example where like it's very

  straightforward that like search ads are

  based on the cost for install and if

  it's a paid app the you know like my

  revenue per install is you know after a

  so whatever you know 70% of whatever the

  purchase price is and I could go in and

  say like I'm will be essentially willing

  to break even say for all of these app

  installs and I try that for a little bit

  but in general that didn't seem to work

  at a very high volume anyway where I

  would it's like I could I would get a

  some amount of install in traffic but it

  didn't seem like I was paying enough for

  those installs that mads are being shown

  very often I better Mayo my terms are

  very low volume or whatever but it

  didn't seem like it one of these things

  that's like great you know I can spend a

  dollar and get a double you know a

  dollar forty back but I can only have

  three or four transactions a day that

  are doing that and so it's like well

  that's that's like net positive but not

  particularly interesting so I had that

  experience on the paid side and then on

  the free side it's so because it's so

  hard to sort of have this feeling of

  what an install is worth and I pursued

  it a little bit but at the end it kind

  of felt like I was just sort of throwing

  money away or at least putting money

  into something that was so hard for me

  to quantify but over time I just sort of

  backed away from it and decided that's


  you know it's okay and I'm probably

  helped a little bit in this that my most

  important apps tend to have reasonably

  good just general search optimization or

  whatever like they show up fairly high

  anyway and so that helps a lot for not

  needing this necessarily because it's

  you know I'm now the second say CAF I

  was previously like the top results for

  a search you know with the now with ads

  I'm the second result because the first

  one the paid result is going to always

  be on top but you know at least being on

  that first screen is probably useful

  enough but yeah I don't know I won't go

  back and forth on how useful they are

  and then in the back of my mind I kind

  of feel like I'm probably missing this

  or in some ways like the funders the

  fundamental like realities of

  advertising where I'm sure if you were a

  just like a typical bit like a larger

  brand advertiser like your goal is not

  necessarily like direct net positive

  cash flow from each advertising by it's

  a much more broad something that you're

  just you're taking this it's like you

  have a marketing budget that whose

  purpose is to just increase brand

  awareness and to increase usage of the

  app as this kind of more general concept

  in some ways even kind of like per

  conversation a couple episodes ago where

  I was talking about why I'm making

  workouts plus plus free where it's like

  there's this general value to being

  well-known beyond the immediate like

  huge sort of return on that investment

  in the less you see within a few days

  and so I go back and forth and if I'm

  just like doing it wrong or not but I

  don't know my experience has been kind

  of mixed and mostly I've kind of backed

  away with it I'm gonna try the new

  basics approach like just to see if

  there's anything interesting there and I

  mean Apple it's giving us $100 free

  credit so like great you know sure I'll

  take that hundred dollars and spend it

  but overall I'm still kind of mixed

  about it I think you know a lot of

  developers didn't give it a fair shot

  when it first launched because we felt

  kind of put out put off by it or kind of

  like you know like it was a trick or an


  thing for Apple to be doing I think we

  have there there's kind of two parts of

  that argument that are that are worth I

  think internalizing now number one is

  that well it's done

  it's it's clearly here to stay they are

  not only keeping it but expanding it so

  I think that that ship has sailed much

  like initially when we brought mad that

  a preview was a thing when the AppStore

  first launched that ship has sailed and

  so has this one so you might as well get

  on board with at least being okay with

  it existing even if you decide not to

  participate or net you know you should

  at least realize that this is if this is

  now a factor in the AppStore that's

  that's probably here to stay um and and

  number two while it does feel pretty bad

  to pay to reach people you were already

  who are already seeking you out so like

  if somebody searches the App Store for

  overcast someone else is that is gonna

  be up top there if they outbid me on

  that keyword that day probably audible

  but someone else is that is gonna be up

  there and it's probably gonna be a

  bigger company than me and I'm gonna

  feel kind of like that's a little bit

  unfair but the reality is that could

  also be a smaller company than may that

  could also be an indie startup who wants

  to get into the same market I'm in and

  wants to have a chance at paying to take

  away some of those searches from me and

  I'm doing the same thing I'm bidding on

  generic terms like podcast but I'm also

  bidding on terms like audible because

  you know I'm trying to get some of the

  same some of that same audience and

  search ads largely have worked for me

  and so it is the kind of thing where you

  can still promote an app for free

  through traditional PR channels and

  marketing channels and you probably

  should but this in a way gives an

  opportunity for small apps to get seen

  in some way and get started and yes of

  course it does take money and therefore

  that will price out a lot of developers

  but if you manage it responsibly and

  conservatively upfront it doesn't have

  to take that much money and you and if

  you if you luck out with your terms and

  if you control it well enough you can

  actually make it profitable per user so

  this leads into I think the first major

  Challenge of using search ads which is

  you have to know what a user is worth to

  you and that's very hard to do if your

  app is not paid upfront if there's any

  more complex business model than paid

  upfront and honestly I don't think paid

  upfront has worked that well in search

  ads because you don't pay per install

  you pay per tap on the ad so the the

  cost per install varies depending on how

  many people tap your ad and then decide

  not to buy it or not to download it and

  so that can obviously vary based on

  things like your screenshot your

  description your ratings but also I

  think if you have a a price there of an

  upfront price I have I would guess that

  causes much lower buy through rates or

  you know conversion rates from from the

  tap to the purchase then a free app

  would and so that's gonna dramatically

  affect your cost right there so you know

  if you're paid upfront I think I think

  it's easier to calculate what an install

  is worth to you but it's also harder to

  get those installs so I'm gonna now

  focus for the rest of my discussion of

  this on free versions on free apps

  upfront because that's what I know

  that's what I have and I think that's

  gonna be mostly what search ads are good

  for because there's just that one less

  big barrier to entry for people to to

  who are learning about it for the first

  time through that ad to then click there

  and download it so the question then

  becomes how do you value a user if your

  app is free upfront so if you have a

  scheme like I do which you know I think

  at least part of this is common if not

  it's not the whole thing where I have

  ads in the app and I get paid for those

  ads and I have an in-app purchase that

  can be optionally bought to remove the

  ads and add some premium stuff but

  mostly it's removing the ads and I you

  know people I think this is common

  enough that I can tell you some lessons

  I've learned from this number one when

  you're trying to figure out what each

  user is worth to you you have to make

  some assumptions that might be

  uncomfortable or cruel or or revealing

  for instance what percentage of people

  who download your app will actually ever

  launch it that's a big it's not a

  hundred percent newsflash if this yeah

  if you've worked aware of this yet you

  know and if you already have some

  analytics in your app you probably are

  no this kind of thing but yeah it's not

  a hundred percent that might you know

  you might get two thirds or eighty

  percent or something like that like it

  might be or half it might not be a very

  very kind ratio of people who actually

  ever launched the app after downloading

  it and then after that you have to then

  work through like if you're valuing the

  user based on the likelihood that

  they're going to buy an in-app purchase

  or whether they're going to see your ads

  or follow through any other kind of

  monetization what other barriers are

  there between getting the app and

  running the app and getting to that

  point where they're making you money do

  they have to create an account

  do they have to sign in do they have to

  you know enter some data and all those

  steps are gonna reduce that percentage

  even further of how many of those

  downloaders actually make it to the part

  of the app that makes you money not to

  mention the fact that once they get you

  know assuming they're actually using it

  then you have the question of what

  percentage of actual users will then by

  the end up purchase or if it's ad based

  what percentage of them will see the ads

  and then you and ads have another

  massive complex guess that you have to

  make or data you have to collect which

  is how long will each person be using my

  app you know will the average user use

  it for a day a year ten years like we so

  those numbers can dramatically change

  what you think a user is worth and I

  would urge you that if you don't know

  this number you know if if you are

  making estimates and guesses on a lot of

  these numbers which you might have to

  you know if you if you don't have really

  good really creepy analytics like I

  don't you have to guess and a lot of

  these numbers or just wait and see the

  results and for me I found that guess

  very conservatively on these numbers

  like every step of the way that might

  reduce the amount of users that get to

  the next stage cut it in half like

  really like be very aggressive like just

  assume half the downloads will result in

  launches reassume that half of the

  launches will result in in accounts

  being created or you know signups

  happening or proceeding to the next step

  and so these numbers starts to start

  said you know divide out and math starts

  catching up to you pretty soon and so

  you have to be very very careful when

  making any kind of assumption

  about what a user is worth to you

  long-term and keep in mind that

  long-term might really be a day for a

  lot for the majority of your users

  it like your average duration of how

  long a user sticks with your app might

  be a day or a week or something pretty

  short so you can't for instance assume

  like eat that you have like six months

  or a year of ad revenue to attribute to

  that download cuz you probably don't the

  nice thing though is we do get a lot of

  those numbers now from app analytics and

  iTunes Connect that's true yeah like at

  least we do have a better starting place

  then like you don't necessarily need the

  creepy analytics for a lot of that kind

  of stuff cuz I'm just thinking like with

  the things that we get from there or

  like we get AG a retention percent which

  is a really useful measure for being

  able to get a sense of like you know so

  once someone has downloaded the app what

  is it like you know what percentage of

  those people will down open it the next

  day and what person will open it a week

  later two weeks later three weeks later

  and they go up to 28 days and so at that

  point you get a sense of like what's

  your retention you look curve looks like

  there and so at least you can kind of

  get you know infer that you know it if

  your retention is kind of flattened out

  by the time you're a month out like we

  won't you've probably gotten that user

  if you're still you know whatever it is

  say you have you ten percent retention

  after a month it's like you probably

  have them for that but even if you just

  capped it at a month and tried to work

  out you know how many sessions does the

  wooly user have you know if they if ten

  percent of them are there after at the

  end of a month and maybe twenty percent

  of them are there after a week you know

  you can kind of do a little bit of math

  but yes it's it's it is a really tricky

  thing I think being conservative here

  makes is by far the better way to air on

  because it's far more important for it

  to just be a reasonable number than it

  is for it to be an exact number exactly

  and all this by the way this is all

  making a pretty big assumption all this

  advice so far that you don't want to

  lose money on each installation but

  there are types of businesses and apps

  and situations where that might not be

  the case you might actually want to lose

  money on or you already might you might

  be willing to lose money on every


  maybe it's only a little bit of money or

  if you're you if you have a plan to make

  it up down the road or if you really

  want to just focus on growth of user

  base right now or if you haven't

  monetized yet if you're like you know a

  start-up that's pre monetization

  strategy and you just want growth you

  know so you might have reasons where you

  don't care about whether you're gonna

  make up the you know make it up properly

  you know on the other side or not or you

  just might be betting for your future in

  bidding you know what I can take a small

  loss on each user right now but some

  percentage of these users will tell

  their friends and spread the app and

  cause my organic installations to go up

  so it's it's very very vague and based

  on a lot of estimates and assumptions

  that this is true of all advertising on

  all sides it's for everything so all

  this to say like you know try to

  quantify some of these numbers try to

  make good estimates to try to have as

  much data as you can to to make good

  decisions here but your decisions might

  not have the same you know needs as

  someone else's anyway before I get into

  specifics of what I do in the App Store

  and with app search ads we are brought

  to you this week by Eero this is another

  kind of ad by Eero never think about

  Wi-Fi again arrow has created the dream

  Wi-Fi setup a fast reliable connection

  throughout your house and even into the

  backyard if you want and now is the best

  time to get on board with arrow as

  they've just released their new super

  slick second generation devices this

  brand new tribe and second generation

  model along with the new arrow beacon

  allows you to build a Wi-Fi system that

  is perfectly tailored to your home the

  new second-generation arrow includes a

  third five eager Hertz radio making it

  twice as fast as before this is to do

  more than ever whether it's your Wi-Fi


  Eero has the power to blanket your

  entire home in fast reliable Wi-Fi the

  base the base unit sits flat on any

  surface just plug into the wall with the

  included power adapter and you're ready

  to connect to your Eero either with

  Ethernet or wirelessly and the new year

  also includes a thread radio which lets

  you connect to low-power devices such as

  locks doorbells and more and the new

  little satellite units are now called

  aero beacons and this is new as well all

  you do is plug it into a wall it looks

  like kind of like a large

  nightlight almost it just it sits flush

  right on the outlet right against the

  wall there's no wire needed plugs right

  in and it even includes a nightlight on

  the bottom to help you out so you can

  plug it in anywhere you have an outlet

  you don't

  need like a table or a surface or a

  cabinet to hide it in these are

  wonderful I love these new beacons and

  the new ero app lets you manage your

  network from the palm of your hand

  and you can also easily create and share

  a guest Network too if you want that and

  if you need any help

  IRA's customer support is amazing you

  can call them and get a Wi-Fi expert on

  the phone in just 30 seconds I've used

  euro and it is by far you know I've used

  other multi-point access point things

  before and first of all you definitely

  want multi-point Wi-Fi but euro is by

  far the easiest and fastest to set up

  I've ever seen it's wonderful and and

  this the fact you don't to run wire

  between the base stations is also just

  incredible so definitely recommend it

  it's it's very very easy to use and

  performs very well so the new euro

  starts at just 399 for one second

  generation ear a base station and two

  beacons and that is everything you will

  need for almost any situation you can

  add more if you want but you probably

  won't need two listeners of this show

  can get free overnight shipping to the

  US or Canada when you head to eurocom

  that's eer Oh calm and use promo code

  radar that's eurocom promo code radar

  for free overnight shipping thank you -

  Iiro for supporting this show so I

  wanted to go into a little bit more

  specifics of how to optimize search ads

  on the App Store now that you know

  roughly what you want to spend per

  installation because you figure that

  during my ad read I guess I don't know

  you figured it out sometimes well assume

  you have that number oh did the math

  yeah yeah well uh flea what you're able

  to spend per installation you know it's

  worth now learning kind of how the

  system works so the main thing with with

  search ads is that it's dependent very

  very heavily on Apple's own App Store

  relevancy engine and as we know for App

  Store search it's not that good it's you

  know it works on a fairly coarse level

  but there's a lot of mismatches a lot of

  misjudgment of relevancy ranking isn't

  so great and this applies therefore to

  the search ads matching algorithm as

  well so you need to do some tailoring

  and you need to know that going into it

  that the relevancy matching is not very

  strong and is you know therefore that

  could work for or against you if you

  know your app might rank weirdly high or

  really low one thing Apple has done

  which i think is a smart move

  is and it's and Phil announced us on

  stage when they announced search ads or

  whatever that was but the idea is that

  even if you bid a high enough price if

  Apple's algorithm deems your app not

  relevant to a query it will not include

  it no matter what no matter what you bid

  and this is good because this prevents

  something like clash of clans bidding on

  the term podcast because it might be

  could be becoming a high-volume term and

  they want to get it on that even though

  their app is not relevant to podcast at

  all you know so that's actually a very

  good prohibition the downside of that is

  that if the search algorithm guess is

  wrong on whether your app is relevant

  you could be on the wrong side of that

  so something to keep in mind there's not

  much you can do about that unfortunately

  but that is nice to know that that's a

  thing when you are picking your keywords

  I have found that it's best to mostly

  just do search match that if you do too

  much keyword selection you can get weird

  results if you do search match you're

  kind of letting Apple do it for you and

  this is actually so when they long

  gesture day long search ads basic that's

  all it is as far as I can as far as I

  can tell it seems like it is a search

  match only ad account is that is that

  right yeah you just give it a budget and

  a maximum cost per install like those

  are the only two numbers you put in

  mm-hmm and then it just does the rest

  it'll do it'll try and find you whatever

  the best opportunities are and you know

  sort of do its own thing but you have no

  visibility into what those terms are or

  anything like that it's just you just

  give it it basically you just tell it

  how much money overall are you gonna

  give to give to this campaign and then

  how much what's the maximum amount that

  you're willing to pay for an install and

  then it just does the rest for you the

  only thing about that I will say that

  then this is why I'm a little bit

  worried about search that's basic is

  that one of the great things about the

  search ads implementation in the

  Advanced Mode that used to be the only

  mode until yesterday is that you can go

  in so run it for like a few weeks you

  know with you know or whatever time you

  want you know run it for a little while

  and then you can go in to the reporting

  areas and you can see if you go to if

  you go within the ad group you go to a

  tab called search terms and it will list

  the term

  people actually use to search your app

  and you can rank that by things like

  conversion rate or average cost per

  installation which they say it's CPA

  cost per action in the advertising world

  but it's a cost per installation and so

  you can you can sort and rank the

  keyword people are actually using that

  Apple is matching to you

  you know for you and you can see like

  you know certain ones are gonna be are

  gonna be better than others in terms of

  what you're actually spending for those

  users what percentage of them are

  actually installing your app etc and if

  you see the keywords like I'll you know

  I'll tell you like some of mine like I

  started out as so if I sort by spend

  which is the the amount of money you

  have spent on it for that time interval

  my first one is podcast which is good

  that is my most expensive one but that's

  also my highest Volume one and I'm

  bidding high on that because I want I

  want to get those people who are

  searching for podcast then there's low

  volume terms just kind of their bucket

  for everything else that's too small to

  matter then I have things like podcast

  app for iPhone great that's what I want

  the next one down is FM transmitter well

  I'm not sure I want that one that's

  that's not that relevant and sure enough

  FM transmitter has a way lower

  conversion rate than the other ones so I

  have to decide like is this actually do

  I actually wanna be paying for people

  who search for FM transmitter and if you

  don't one thing you can do in searched

  in search ads advanced is you can you

  can go through this list you can check

  off whichever ones you don't think are

  worth bidding on that are not relevant

  enough and you can say add as negative

  keywords and if you say exact match then

  which you probably should do then it'll

  it will not put you on those bids and

  that I have found is a necessary part if

  you're gonna do search match if you're

  not gonna just put on the exact key

  which if you're gonna do any search

  matching at all on the App Store you

  must go in and add as negative keywords

  the things that are not relevant to your

  app that are costing you a lot of money

  and my might because you know because

  Apple's relevancy engine isn't that

  great you get a lot of these and some of

  them you actually might want so like one

  of the very common ones that I get is

  things like music player without Wi-Fi

  free music downloader offline music

  player FM radio offline

  and I'm like well actually those are

  people who want to listen to stuff

  offline some of which is radio so I

  think I actually might want to keep

  bidding on a lot of those whereas one of

  them over watch I don't want to bid on

  because that's a video game there's

  nothing to do with my app and it's being

  matched by some kind of linguistic

  coincidence and I definitely don't want

  to bet on overwatch terms so I

  definitely add that as a negative

  keyword but again keep in mind that

  you're dealing with a fairly low

  sophistication search algorithm and

  relevancy algorithm so be careful one of

  the reasons I said only add negative

  keywords as exact match is that earlier

  in the summer I had a problem where I

  had added too many things to negative

  keywords as like regular like broad

  match like relevancy match thinking that

  was what I think to do and I started

  getting matched on very very few things

  I'm all of my numbers just plummeted and

  I contacted the search as team and

  they're they're very nice there's actual

  humans there who will like help you out

  and they ran amongst numbers and they

  recommended that I like remove some of

  those and just switch to a more

  relevancy based approach which I did and

  and so I would recommend only using

  negative keywords and exact match mode

  to avoid that situation anyway

  I end up you know coming out for for a

  long time until the last couple of

  months I have come out ahead in what I

  believe a user to my app is worth versus

  what I'm paying per install from the

  search ads and the search ads have made

  up approximately 10 to 15% of my weekly

  installs and so that's not bad that's a

  lot yeah exactly and that's 10 to 15

  percent of my installs per week so you

  know that that increases over time and

  like those users are you know sticking

  with the app I think I don't know there

  actually is an API somewhere in

  somewhere in iOS there's an API where

  you can attribute installations to your

  app to whether they came from search ads

  or not and I think even what keywords

  they came from and then you can tell in

  your app analytics like what those

  people are worth to you I have not done

  that that seems a little bit too fiddly

  for me I don't think I really care that

  deeply and it's a little bit creepy I

  think but but so I haven't done that but

  you can do that and you can get into you

  know deeper measurements but

  I have found for the most part this this

  has been worth it for me to do for a

  while unfortunately over the last couple

  of months the podcast related keywords

  have really heated up

  thanks audible from like big companies

  thanks audible who who don't seem to

  have a limit to their budget thanks

  audible and so I I'm losing a lot more

  bids and I've had to increase my bids if

  I want to keep winning them and I'm now

  pushing to the point where I'm losing

  them a little bit of money on each

  install that comes from a search ad and

  so so far I've decided well I I kind of

  want to wait and see and watch how these

  numbers go I I kind of don't want to

  give up that 15% so maybe I'm okay

  losing a bit of money on those people

  because then the other 85% of my users

  will make up for it so I'm not entirely

  sure I'm gonna do about this long-term

  yet but the only reason I've been able

  to keep costs under control at all is

  because when going in and negative key

  wording the searches that are providing

  me with really bad conversion rates and

  really high cost per install and because

  search ads basic does not offer that

  ability I do hesitate to recommend it

  because I don't think you're gonna get a

  lot of control over that kind of thing

  and Apple's search relevancy engine is

  not good enough that I would trust us to

  do it without any intervention for me

  what do you think yeah I mean I think

  I'm intrigued to go to use the basic

  system but mostly insofar as it's an int

  because like because it's the only input

  I give is with the maximum amount I'm

  willing to pay for installation is I

  will just put that number low enough

  that I feel comfortable with it and

  understand that like your approach of

  the highly tuned like going in there and

  looking at keyword by keyword like

  dialing it in you are going to end up

  with a better result for the same like

  cost per install but it takes a lot of

  work and thoughtfulness and so on versus

  the approach you're just saying it's

  like I'll put that number low and if the

  number is so low that it's not getting

  high volume okay like that's fine I

  didn't have to do much for it but if I

  am getting installs from that great and

  so I kind of view it as a you

  over you what you're doing makes perfect

  sense I think it's the ideal way if you

  if this is something that you want to

  spend time on if you want to in sort of

  have it be a focus if not if you're sort

  of like me and it just like this is an

  interesting kind of more amusement it's

  an interest it's a really convenient way

  to just kind of like a set it up set it

  at a number that you're comfortable with

  and then just see what happens and

  that's how I expect to proceed from here

  well best of luck I do I strongly

  encourage everyone to use that free

  credit look you might as well they're

  giving you a free credit so use it and

  see how it goes if you want to be she

  talked about it wait until after

  everyone else uses theirs then use yours

  sure then the price will be lower but

  anyway we're out of time this week I

  hope this was helpful to you

  thanks to listening except thanks

  listening everybody and we will talk to

  you next week bye