Instagram is incredibly polished. Notice how fast it is? As you're selecting an image, they're already uploading it to their servers. That is critical to maintaining and growing their user base. I go to Instagram knowing I can upload a nice looking photo quickly. The images load incredibly fast as well. Its simplicity is deceptively robust.
This is not recreated by a developer billing at $50.00 an hour.
So, if you're looking to have an app built, and you're reading this post, just know it's pretty inaccurate. You get what you pay for.
Contract iOS engineer here, based out of Colorado. My company bills $150-$200/hour for my time.
> Its simplicity is deceptively robust.
I've met some of the Instagram iOS team. Part of the beautiful simplicity comes from the fact that FB can afford to pay ~5-7 iOS engineers top salary to work on improving the app full time. And that's just the iOS app.
I agree with you that $50 is stupidly low. I also agree with you that the estimates are laughable for making an actual copy of Instagram.
BUTTTTTTT. They are very clear in the article about calling it an MVP of something like Instagram. It's the difference between making the Twitter-like app in a tutorial and making Twitter.
This is a pretty good example of someone making an estimate for an MVP level contract app. Obviously it's what they do for a living so that's not surprising. I think it's a good reference for anyone else trying to spec something. If you aren't at this level of detail, you don't have enough detail.
The big problem with a spec like this is that the app you will get is not going to be anywhere near the level of quality of an app like Instagram. You have to set customer expectations or you will spend a huge amount of time litigating whether or not the app is 'done'.
> The big problem with a spec like this is that the app you will get is not going to be anywhere near the level of quality of an app like Instagram. You have to set customer expectations or you will spend a huge amount of time litigating whether or not the app is 'done'.
This. In my experience of building apps under contract, the concept of the MVP is not very well understood outside of our tech industry echo chamber. We regularly have clients who know the term and "want an MVP" but when it comes down to release act completely contrarily (mostly trying to delay releases for additional features, more polish, etc). Client education and carefully setting expectations are key in this line of work.
This estimate also complete ignores billable time for project management and testing.
Edit: And design, if you want your app to look even remotely different from Instagram. And devops to stand it up in a scalable way. And maintenance / service agreements to keep the thing running...
You may be wrong. One of the founder of Instagram has said he wasn't a very good coder when they built the product. He further said some of that old code was still inside the product.
It is 2016. Development work is pretty much copy and paste instead of rocket science.
Based on this article http://www.recode.net/2016/6/8/11883518/app-boom-over-snapch..., the whole app ecosystem is dying, and less and less people are going to explore new apps. In the short future, people will stop paying the insane amount of money to developers to build the apps nobody is using.
For app development, the hard part has already been done through well-written lib, and the developers mainly write some trivial business logic to put these components together.
When instagram is made there was no library of making actual logic. You should not rely on libraries if you want to make instagram you should rely on people who can make that kind of libraries
I would be "impressed" with your productivity if you write an authentication, message and notification system from scratch over twice, and each time takes over 40 hours.
I think you missed the point of the post which is not what it'd cost to build Instagram, but to build a MVP camera/album + backend, similar to the basic functionality that Instagram offers, without the level of polish.
As for the $50 rate, it kind of depends. $50 an hour, at a 40 hour workweek year round puts you at roughly $100k a year, pretty reasonable salary I think at an experience level between jr and sr. And in a startup you can even do a bit of equity financing and pay a lower salary, in terms of pure cashflow $50 could even be on the high end in highly bootstrapped startups. But that's if you employ someone.
If you outsource it to a decent dev team, then no way, you always pay a substantial multiple. For example I recall at my job as a student doing administrative work at a big corp, I had a $15 an hour wage and billed the client $120 an hour, no joke. My boss had a $50 an hour salary and billed $400. $150 and up per hour is pretty typical afaik.
You make it sound like $50/hr devs are stupid and not able to devise such a clever trick but reality is not like that.
Sure there is a difference but it's not in your example, the pre-uploading feature was hardly there in the first app versions and it's not something that some smart iOS dev invents on a whim and bakes into an app, it's a deliberate feature requiring collaboration between several teams.
Most of all, features like that are created because there are (amazing) devs whose sole job is to make the app better, and they've been doing it for what, 6 years now.
Or, granted you know how to quickly setup the push certs, and deploy a simple docker container with parse-server to AWS Beanstalk, you could launch a MVP (iOS, Android) in one afternoon, for $49.
The hour estimates are off by at least an order of magnitude.
Not included is the infrastructure necessary to ship something like Instagram (testing, monitoring, deployment, beta, etc), even if you leave the massive scale of Instagram out of the picture.
Much more significant is that even the relatively rudimentary 2010-era Instagram was the result of thousands of hours of polish and refinement after the initial product was built and released to the public.
There are zero products that go directly from mockups to success. It's relatively easy to knock off something that already exists and is proven to work.
> There are zero products that go directly from mockups to success. It's relatively easy to knock off something that already exists and is proven to work.
Building a knockoff is a relatively linear process. The challenges have already been solved and the work that needs to be done is clear to all involved.
Creating a new product is significantly less linear and a lot of budget gets burned because of that. The time of engineers, designers, project managers and QA all cost money. Aligning them to the same (non-obvious) vision therefore costs money. There are additional real costs every time the plan changes even slightly (scope creep, customer feedback, founder waffling, etc).
"No battle plan ever survives contact with the enemy"
- Helmuth von Moltke
I found that very relatable as we recently made an app close to instagram in UI sense and with couple of features plus minus in react native . The app(react native) and backend(express) were written by me and including everything took easily ~10-15 days more as predicted by the author .
For 10,000 can you imagine someone building Instagram as it was available on the day it was released? How about 2 months after it was released. 2 years?
Sure, there's a point where the scale of Instagram tipped over and they had to hire more than 2 engineers, but that point was 30 million people a day, and even then it was still recognizable as "Instagram" the service and the app.
They don't have to set out to duplicate the billion dollar company from day one in order to achieve their goals. Simply being a knockoff from the early days would be good enough.
Why does it take over 40 hours to implement features, which have been implemented for millions of times and have existing solutions and code in the Internet. Authorization (Social and Email) seems to be one or two hours of work with existing solutions.
And you think you can build a great integrated experience that just works with multiple platforms in "one or two hours of work"? Am I right in assuming that you never built systems like this?
I agree, according to the article, the front end of "email authorization" takes the most time to do. For an MVP a login system really just has to work, and then you can personalize it more as you go. Most of the time for an MVP should be dedicated to the main features of app. (In this case uploading pictures, news feed, etc).
For my app we used Auth0 to quickly get login working before we had to present it. It took about total 5 hours to set up and test. Since then we have built our own login system and made sleek looking create/login screens; which in total took about ~15 hours.
Most people have an idea for a great app that they keep playing in their head. But these ideas never come live because either they don't have the expertise to build it themselves or they don't want to drop $12K just to see if users want it. How much will you be willing to pay for the MVP of you idea? (I have lined up a reliable team who can deliver most MVP app in 6 weeks at around $3800. Need validation for the price point).
I want an app that aggregates commercial loan data from across the industry, which is un-standardized, and creates a deep learning platform with a public api that can tell a customer what his individual risk is for both sides of the party in regards to underwriting, processing, and securitization. Based on real time market data and regional inputs. Also taking into account who I've done business with before and relationships.
The backend estimation is laughable. I don't think the author really understand what he/she is talking about. For example, search. Having a search functionality is one thing, having a usable search is completely a different thing.
Clearly the article is by some dev shop and anybody who has worked on an app with the scale of Instagram, or even a fraction of Instagram's scale, knows that this entire article is misleading. If you wanted to actually create Instagram, as it is Today, it would be an incredibly large undertaking.
The estimates are for a "functional prototype", which is, the implementation of ONLY functional requirements. Non-functional requirements are not mentioned here, and in the case of Instagram, there are many.
Then, since Instagram is a service, you need some sort of operations team to manage and monitor the infrastructure. You will also need some sort of testing (quality assurance, performance and scalability, security...). Then you will need some degree of organization to accomplish this.
Additionally you will need some visibility and minimal critical mass of users to get the ball rolling (aka user acquisition). That is of course, not development, but this is one of the most expensive items, to the point that for some apps, development costs are marginal compared to user acquisition costs.
You might get a software with the Instagram user-facing features, but don't expect to be able to handle the volume of traffic that Instagram does, at the level of security they do. That last part is actually a key aspect of the business model that I couldn't found anywhere in the article.
Thank you! i'm so glad someone did a post like this, maybe finally we can stop having people say things like 'oh i can build an MVP of Twitter / Instagram in a weekend, its just a photo sharing app'. No. There are a LOT of little features that need to be considered and they all add up (authentication / settings / notifications / mentioning etc etc).
59 comments
[ 4.2 ms ] story [ 238 ms ] threadInstagram is incredibly polished. Notice how fast it is? As you're selecting an image, they're already uploading it to their servers. That is critical to maintaining and growing their user base. I go to Instagram knowing I can upload a nice looking photo quickly. The images load incredibly fast as well. Its simplicity is deceptively robust.
This is not recreated by a developer billing at $50.00 an hour.
So, if you're looking to have an app built, and you're reading this post, just know it's pretty inaccurate. You get what you pay for.
> Its simplicity is deceptively robust.
I've met some of the Instagram iOS team. Part of the beautiful simplicity comes from the fact that FB can afford to pay ~5-7 iOS engineers top salary to work on improving the app full time. And that's just the iOS app.
And as the product matures the develop-test-deploy cycle gets slower
BUTTTTTTT. They are very clear in the article about calling it an MVP of something like Instagram. It's the difference between making the Twitter-like app in a tutorial and making Twitter.
This is a pretty good example of someone making an estimate for an MVP level contract app. Obviously it's what they do for a living so that's not surprising. I think it's a good reference for anyone else trying to spec something. If you aren't at this level of detail, you don't have enough detail.
The big problem with a spec like this is that the app you will get is not going to be anywhere near the level of quality of an app like Instagram. You have to set customer expectations or you will spend a huge amount of time litigating whether or not the app is 'done'.
This. In my experience of building apps under contract, the concept of the MVP is not very well understood outside of our tech industry echo chamber. We regularly have clients who know the term and "want an MVP" but when it comes down to release act completely contrarily (mostly trying to delay releases for additional features, more polish, etc). Client education and carefully setting expectations are key in this line of work.
This estimate also complete ignores billable time for project management and testing.
Edit: And design, if you want your app to look even remotely different from Instagram. And devops to stand it up in a scalable way. And maintenance / service agreements to keep the thing running...
You don't create software, do you?
As for the $50 rate, it kind of depends. $50 an hour, at a 40 hour workweek year round puts you at roughly $100k a year, pretty reasonable salary I think at an experience level between jr and sr. And in a startup you can even do a bit of equity financing and pay a lower salary, in terms of pure cashflow $50 could even be on the high end in highly bootstrapped startups. But that's if you employ someone.
If you outsource it to a decent dev team, then no way, you always pay a substantial multiple. For example I recall at my job as a student doing administrative work at a big corp, I had a $15 an hour wage and billed the client $120 an hour, no joke. My boss had a $50 an hour salary and billed $400. $150 and up per hour is pretty typical afaik.
Sure there is a difference but it's not in your example, the pre-uploading feature was hardly there in the first app versions and it's not something that some smart iOS dev invents on a whim and bakes into an app, it's a deliberate feature requiring collaboration between several teams.
Most of all, features like that are created because there are (amazing) devs whose sole job is to make the app better, and they've been doing it for what, 6 years now.
https://market.ionic.io/starters/photogram-ionic-instagram-c...
IMHO MVP means a product that users can use, but not full-blown scalable platform.
Not included is the infrastructure necessary to ship something like Instagram (testing, monitoring, deployment, beta, etc), even if you leave the massive scale of Instagram out of the picture.
Much more significant is that even the relatively rudimentary 2010-era Instagram was the result of thousands of hours of polish and refinement after the initial product was built and released to the public.
There are zero products that go directly from mockups to success. It's relatively easy to knock off something that already exists and is proven to work.
Building a knockoff is a relatively linear process. The challenges have already been solved and the work that needs to be done is clear to all involved.
Creating a new product is significantly less linear and a lot of budget gets burned because of that. The time of engineers, designers, project managers and QA all cost money. Aligning them to the same (non-obvious) vision therefore costs money. There are additional real costs every time the plan changes even slightly (scope creep, customer feedback, founder waffling, etc).
"No battle plan ever survives contact with the enemy" - Helmuth von Moltke
Just hire a 10X programmer for $50/hr.
You're not going to scale to an 80 million pics per day service[0] on a $10,000.00 backend.
[0] Instagram is not an app, but a service.
Sure, there's a point where the scale of Instagram tipped over and they had to hire more than 2 engineers, but that point was 30 million people a day, and even then it was still recognizable as "Instagram" the service and the app.
They don't have to set out to duplicate the billion dollar company from day one in order to achieve their goals. Simply being a knockoff from the early days would be good enough.
You don't even need to write code if you use https://github.com/rethinkdb/horizon, firebase, or similar services.
For client side, you can reuse previously written code for authorization logic and just write user interface based on requirement.
40 hours are too much for such work
For my app we used Auth0 to quickly get login working before we had to present it. It took about total 5 hours to set up and test. Since then we have built our own login system and made sleek looking create/login screens; which in total took about ~15 hours.
How much would that MVP cost me?
(hint: $15 million)
The last few app requests were: Tinder for shopping, QR code generator and scanner, chatroom app etc.
Then, since Instagram is a service, you need some sort of operations team to manage and monitor the infrastructure. You will also need some sort of testing (quality assurance, performance and scalability, security...). Then you will need some degree of organization to accomplish this.
Additionally you will need some visibility and minimal critical mass of users to get the ball rolling (aka user acquisition). That is of course, not development, but this is one of the most expensive items, to the point that for some apps, development costs are marginal compared to user acquisition costs.
You might get a software with the Instagram user-facing features, but don't expect to be able to handle the volume of traffic that Instagram does, at the level of security they do. That last part is actually a key aspect of the business model that I couldn't found anywhere in the article.
They say hindsight is 20-20, this guy is going at it with a telescope..
http://www.slideshare.net/InfoQ/how-a-small-team-scales-inst...