To Increase App Subscriptions You Need to Look At This Data with Lior Barak

I see and when you were caught in talking about this is a client that I currently have right now where you're like okay look at the installs look at what they're doing put them into cohorts and it figure out like you know if you want them to subscribe how long it tastes to subscribe and then work backwards right like if it takes them you know seven days for instance let's think about ways that we can drip campaigns or like well how would you use that data to say okay you are I know my customers you know they retain they subscribe within the first seven days or they don't how would I use that data then - like what kind of things should I be doing with that data so in the first point of it is actually to understand when your revenue event is happening so when is your conversion happening the second phase will be basically to do a retargeting campaigns so as I see it I'm quite I'm building on the fact that

I'm saying that most users of Earth gonna delete the atma stuck in the first 30 days if they haven't returned at least twice in the first seven days this is this is data that we look quite often on subscription-based apps for example and this is quite interesting because this means that if in the cell in the first seven days you didn't have at least my a visit from the user on top of the install event already you should check why and how can you trigger it again you can start already there and if we after 30 days you know that most few users uninstall direct you need to do again and you see campaigns actually to to attract them back to install the app again but you can say that in the first 30 days you don't attract this user to re-install the app you'll give them the time and then after 30 days or after 60 days off in 90 days you're actually starting to target them against which installs until then you're

trying to retarget them to open their and and you're building some strategy basically that's saying how you should spread your money because we have a limited budget we all have been good budget it's not that like we have millions sitting in our bank account just waiting for us to spend them and and the idea it is to take this data and actually understand how the users react and based on their reaction to decide when and how much money do I spend because there is a certain amount that you can spend on the user right because you know usually a revenue for users are going to be $15 and if I gonna spend now $100 on him does it make sense yeah do you help your client with that with the retargeting campaigns or do you recommend somebody else we don't do campaigns at all we have been in automation of campaigns so we're doing the automations processes but we are not coming from the marketing perspective we're coming from the data

perspective we're sitting just in the middle between the engineers let's say and the marketing on the other side and we are the one we are the arm that operating both of them to basically drive better results because on the one side you need the engineers to be databases you need the engineers to build the automation and on the other side you need the marketers to keep doing campaigns so the machines can learn and can readjust and you can basically improving your business yeah the I love that thing what you talked about retargeting because I remember I did a article about this but Iago from pique you know peak brain training they do a phenomenal job of retargeting that's what you talked about like hey we know where those revenue points are kind of like what you talked about and they use retargeting campaigns because they know if they can bring people back in with that first seven days they're more

likely to subscribe and so they really rely heavily on retargeting campaigns and showing the most popular games and so if you have more popular content making sure you utilize that in your ads to bring them back into your app anything you want to add on that Lior know exactly you need you need to understand is that they need to have these data to actually make this decision that while you're just gonna do it blindly it's it's a shame it's so true the other thing I want to talk about is when you're working with you to bigger clients who are spending let's say tens of thousands of dollars maybe $1,000 a day and you're talking about Rojas now you're talking about looking getting data from Google UAC and Facebook what are you then now looking at because there's so much data points now so what we as the bigger the client is the more narrowing we doing their data so there is a very important fuel

in what we do and what we say is that you cannot have more than three daily KPIs to optimize your campaigns there are only three you can use this term Maxim you can do and what we basically doing is that the understanding that we cannot tell them okay you just use now our Y we're just using some certain metric on the final of the session so we're doing like a campaign ranking which this is basically coming from a data scientist it building a model that says based on these events these are the chances that this campaign is a successful campaign compared to other campaigns that we have and we're using all this data but we will minimize it so we're taking recompressing tons of points on the way into one number but saying this is eight point seven if you remember we talked about it in the last interview we had we talked about the indexing and this is exactly what we're trying to do we're trying to

give people the indication based on a lot of events on the session what is actually the the Wharf of the session itself oh the the user they've just brought yeah if you guys want to listen to that I'll link up to that episode as well but six nine six so if you want to just checked out the previous episode you can definitely do that Dean what I want to move on is when you say let's become data-driven right like what does that really mean because I think you meant she did - it's we look at all this data it's not so much about collecting it just like what the hell do we do with all this data exactly so when we're talking about data-driven it means that we understanding the data we understanding the funnel of the data and we understanding what we should use so this is another mistake that I'm seeing quite often is that people don't know which measures to you so we have clicks that coming from Facebook and Google for

example but we also have click information it's happening coming from a just and from app spider and now some of the people were just mixing them all together they're taking the impressions from Facebook and Google and the connects from a just and they did trying to do a city hour from that and I ranking any score and it this is not how it should be it should be much more simplified and it should be one number they're taking you all the way and then if attacking impression we should use the clicks that coming actually from the provider from the ad partner while if you want to see on clicks that come from just there is always a difference there is always like some discrepancy in the numbers and we cannot use it so to become a really data-driven it means that you understand the difference between the data sources and why is it happening and how you can fix it or how or which data you should use and then at the end

of it it's neat to be actionable what is the action I'm driving out of this data I'm collecting you can collect a lot but do you actually be driving any actions if you don't drive actions just keep it in your graveyard and one day you can call it and start using it but bury it for now and don't don't look at it it just gonna mix you up when you say like focus on the three daily KPIs is that just client depended no not at all it's everybody everybody should focus up to three KPIs there should not be more than 3k yes because we cannot make decisions why are the KPI is dependent on the client we need across the board do you say hey you should always look at installs you should always look at revenue points are there like specific ones a KPI is that you should always look at or is it really client dependent so it's client dependent most of them coming at the end of the day - yes I want to have installs but then to be

honest I don't think that installs is the number that can help you on a daily base this is something you need to make you only weekly base because there is no action that going to be driven if you're gonna see on a daily basis that your instant numbers are going down up because if we're looking at the days of the week right so usually Fridays are quite low then Saturday Sunday there is a you chump so that usually at Sundays there is a huge peak of installs because people sitting at home they're bored they're searching the app and they're downloading something and and we cannot say that install is the number that you need to look at in the database if you know if you ask me what you should look on a daily basis this should be actually the campaign a ranking score it should be what is the revenues that you were driving it should be how much money you spend this is for me the three KPI is more or less that I will put there and I

know the Troy is changing quite often but at least you understanding okay how am I breaking third day of spending goals how is the revenue goes and if there is a correlation that you will immediately see it and it is something that you can actually act right because if you're reducing your spending and you see that the revenues go down then there is some that you're doing wrong so you need to figure it out or if there is a huge increase in your cost but the revenues are not going up or the actually declining this is again a great mark right this is actions that you can do on a daily base while looking at installs it's something that it's not gonna help you in a database it's maybe a weekly base to know okay I'm doing a great job I was driving 50,000 installs this week okay you drive me crazy too looking at Italy yeah we all want climb we're trying to really get them above a thousand it's

just like you know I'm constantly looking at it every day it's like how do we go up there I don't go up there yeah you're right it does drive me a little bit crazy so if you say okay look when like I think I know my audience a little bit where they just might not be spending as much so like okay I have this cost per install data but like it's not that important to me because I might be spending a hundred dollars here how many dollars there it's just a different test different stuff so if I don't have that cost per install is there another metric you like okay we'll replace that with installs maybe you can because this with installs of course but for me as I see it we we should not be driven by emotion and once you're looking at in sauce this is emotional feed this is not any more logical thing when you're looking at revenue when you're looking I don't know click amount of ads viewed this is something that they can do also

you Dave on the organic is not spending any money this is less emotional because there is a logic behind it that says I need to have an average fifteen or twenty five ads viewed every day to make actually actually profit and there is a logic behind it and this is something that needs to be very important on a daily basis we put the motion on the side and we just going based on logic this is a very important part of it I like how you said a motion about it too so true as you can tell I was very it's all numbers I'll er like tell me about the new product that's gonna really help people analyze your data better so what we basically developed and we were talking with a lot of clients in the past months and what we found out it's actually that the problem is in understanding how to use it other correctly and what we've done is we build a tool that going and collecting the data for you from the different

partners and then it gives you in a way of a funnel basically that we created we call it the data discovery workshop it's basically it's building it in three sections and in each section it's reducing a little bit more than mound of information you can get the end of it well we going on the first section we allowing them to see all the information that they have they can ask whatever questions they want so how many installs do I have what are my cows what are my walls and so on and so on then in the second section we actually tell them okay out of these questions what are the measures that you think you need what are the master what is the nice to have and we limit them modeled it so if they have 50 or 60 measures before suddenly they have a place the place holder now for 2500 so they need to be very very thoughtful about what they're using and this is basically how the data like the tables themselves

going to be created and then the first step is actually the visualization so how we taking your measures we creating kpi's but here also we limit you so we're having three tables that we're going to create for you three dashboard sorry but you're going to create for you one is going to be a database which this is the action driven table there you basically you're coming you go by logic this is the companion to kill this the companion to improve this great is looking great then on a weekly basis you can do a little bit more research to support you to make you feel good you know you tap your your shoulder and say I did a great job this day and then the monthly one is more evolved you so it's more for the teams to go into a deep dive and really understand what happened what went wrong and what should be fixed for the map to come because when we're looking at the budget we have a fixed budget most of the time for the entire

year and then it breaks into months and weeks but to change stuff too often not going to be healthy for your campaign it's not going to be healthy for the clients as well so we're trying to do it actually once a month and then based on you to connect decisions I should have invested more money into Facebook because actually snapchat was a was the money sucker this man and I spend a lot of money there and this is basically the idea of what we build there so we having this customer to arrive two to three different dashboards that they can easily understand what they need to do next I like that now like that how you limited so I'm gonna imagine Lior that when you work with clients more on a hand-holding basis or an on product basis that you're doing pretty much the same things like hey you know let's link up all your analytics and all your data so that we can pull it all and then tell me the key metrics that you want to

focus on and we're gonna help you sort of analyze all that stuff exactly so we're going to clients and we have them basically how how they need to build their data to actually make more automated decisions most of our time we're seeing around 60% or 70% of the days on campaign creation and campaign optimization and we should not waste this time it's a waste of time that can be automated quite easily if you're going to go into right tools and this is exactly where we come into the picture we have in them to understand what are the right tools for them you know I don't know if you feel this way but when I talk to clients sometimes like there's always the same ASO tips or like same different like growth hacks that I'm going over there I'm shocked that not everybody knows about when you're talking to your your clients or maybe potential clients are there things that you're like man this is something that I

constantly repeat and so I'm just gonna say it on Steve's video exactly just wait to this video and be like aid go watch this video video before a call don't use excellent that should be the number one issue that I'm facing with most of my client when I arriving even if they have a state of the art reporting system which is working perfectly fine like the blow or low key or whatever power bi sitting lead they still downloading the data and working everything exit because they don't trust the numbers and this is also things that people should understand they need to trust the numbers and you should not start working affects it because Excel just causing more issues it's a time waster on the one side on the other one if you done the vlookup wrongly everything is screwed go make decisions based on these laughter words what do you use then I mean why I got this dump data dump like what am I suppose but it this is exactly why we

build this product we're saying you have a data dump but we actually construct it for you in the right way so you don't need to go to the dump you don't need to with any shitty data or any any issues with your data you're just getting a product and this is how it should be this is most of the clients if they don't have a data engineer on board this is they should find the tool that's going to automate it for them so they can make decisions in a easy way and this is this is the most important part and the ones who ever thought engineer they should stick with the data engineer pull him in this year and make him work and actually make something out of it you know so that we have the data let's actually make something I can use and this is this is a very important message I think

you