There's a problem with speed at my business. Tell me how you solved yours!!!
Abdurrahman Yiğit
19 replies
Hey! I have my own business and we build custom mobile apps from scratch. However, a friend of mine told me that the development process takes longer than normal.
My question to you is how do you have a solution to shorten this process? Please consider that we don't want to sacrifice code quality while shortening the time.
Replies
Vaibhav@vaibhavdwivedi
Start by looking at areas where things can be further optimized.
Share
@vaibhavdwivedi I believe I can optimize repeated tasks. I thought of using a boilerplate for this. What do you think? How do you automate repeated steps? Is it a good choice to use a boilerplate?
At the same time, I am thinking of modularising some commonly used pieces of code and using them quickly in different projects
The only question on my mind is:
Since each application needs different things, how possible is it to turn these features into modules?
@apodebana You can surely use boilerplates if you know it inside out. Otherwise, you might find yourself spending a lot of time just learning about it.
Also, I personally believe you should not jump into modularisation in the early stages. Just my opinion. As long as you know where things are and how you can optimize them, just go for it.
Optim'i
Hi,
What's your friend background ? Does he have all context ? Be caution with external advice on how fast your development should go. A good start can be to ask your team how they feel the development is going, what's slowing them down ...
Now if there's really an issue, you need to identify where it comes from. Is it product, design, software engineering, all of them ? Sometimes the fix can be really easy, like improving a few process, defining the scope better or things like that. Sometimes it can be more difficult, or even be an issue with your team member's skills.
But we'll need more details to help
Optim'i
@apodebana i dit built two boilerplate like this, with the same logic you have
but in the end, for most project I did it ended up making things more complicated because they all have core differences that would have made another implementation more efficient
And those task aren’t really long after all once you know how to do them
That’s why I stopped using boilerplate and I’m now using more things like modules
But it also depends what you’re selling. I was selling highly custom SaaS development for startup, with a need to quickly pivot and iterate the first years, so they all had their specific needs. If you’re always selling the same kind of project then it might make sense, like some people use no code for landing pages
@melmouk First of all, thank you for trying to help.
Yes, my friend has most of the information in general.
I think the main problem we are experiencing is in the field of software.
We cannot do repeated tasks in a more optimised way. For example, starting a new project. I also thought of using a boilerplate for this.
Or features that we build a lot. For example, real-time messaging. I'm thinking of making this a module that can be used in different applications.
What do you think about this idea? Before I start testing (because we have limited time and budget), I would like to get ideas from experienced people in this field about whether this idea can work or not.
The only question on my mind is:
Since each application needs different things, how possible is it to turn these features into modules?
Optim'i
@apodebana I'm not convinced about using a boilerplate, I've tried this in my agency at the time but I found that it's better to start from scratch with the specificity of the project in mind. Except if you're really always building very similar projects.
The module idea seems like a good one to me, I have a few modules / helpers / services I reuse across all my project as they're always more or less the same feature.
I've done it in multiple ways :
- Use case with repository that are framework agnostic to guide the implementation (clean arch inspiration)
- Modules back + module front without CSS or without HTML but with all the methods and helper that I can import into my apps
@melmouk
Ye your right.
What was the problem with the boilerplate you tried to use before?
By boilerplate I meant something like this:
Implementations of commonly used features such as:
Onboarding, RevenueCat, Localization, Stripe, Firebase Cloud Messaging (Push Notification), Analytics etc.
Software architecture and folder structure:
State Management Structure, Data-Flow Structure, API layer etc.
Since those features are very common I thought we can build a boilerplate and maybe even sell it in the future.
As an agency owner would you use that kind of a boilerplate? If no, why would you not?
Don't worry about the learning curve tho. I have something on my mind to fix that.
It is not an advertisement, Im tryna see something :D
Elon Jump
Ever used ‘speed’ supplements? Dunno occasionally works for me.
One thing that helped speed things up for us was optimizing our database queries. We analyzed slow queries, added indexes where needed, and restructured some of the data to avoid complex joins. Also took a hard look at our backend code to identify any bottlenecks or inefficient algorithms that could be improved. Little optimizations can really add up! What part of your stack seems to be the speed pain point?
@rachelmariethompson Yes you are totally correct. We also had the same issue in our backend and servers. After optimizing queries and server hardware we were good to go.
But right know we have a different problem.
There are a lot of repetitive tasks such as implementing Stripe.
I thought maybe we can build a ready-to-use module for that and implement that module to any app we want.
It also brings the idea of selling that kind of modules in the future to my mind.
The main goal is to make commonly used features 'modules' and use them. The only problem might be customizability.
Would you use that kind of modules that brings some features to your app with easy setup and documentation?
It is not an advertisement, Im tryna see something :D
Absolutely! Streamlining workflows was huge for us. We mapped out all our core processes end-to-end, identified bottlenecks and redundancies, and ruthlessly simplified. Automation was also key - any repetitive task that could be scripted or templated, we did it. Saved a ton of time. The other big one was improving communication - we set clearer expectations around timelines and deliverables, and invested in better collaboration tools so there was less back-and-forth. What part of your process seems to be the main speed bump? Break it down and optimize each component!
To solve speed issues, we streamlined workflows by automating repetitive tasks and prioritising clear communication. Empowering the team to make quick decisions also helped reduce bottlenecks and boost overall efficiency.
@nitesh_jamod I'm having trouble with the step of automating repeated tasks. I thought of using a boilerplate for this. What do you think? How do you automate repeated steps? Is it a good choice to use a boilerplate?
At the same time, I am thinking of modularising some commonly used pieces of code and using them quickly in different projects
The only question on my mind is:
Since each application needs different things, how possible is it to turn these features into modules?
@apodebana Using a boilerplate for automating repeated tasks can be a great starting point, especially for tasks that follow similar patterns. Modularizing commonly used code is also smart—it allows for flexibility across projects. You can design the modules to be adaptable, so they work for different applications by tweaking configurations or parameters.
I started using tools that help manage and prioritize tasks more effectively. This way, we focused on high-priority items and got things done faster.
I had the same issue, and what helped was upgrading our internet connection. It made a huge difference in speed and overall efficiency.
@stacey_con_nolly It's good that you found that the problem is caused by the internet speed, but I don't think that would apply in our scenario. I think it's more of a problem with our software architecture