AWS Honeycode is going to make Business App software cheaper for everyone and raise the bar on software quality.
Coming from an App-Dev background the thought of no-code or low-code solutions has always raised mixed feelings. Let’s not get started on the damage that Visual Basic did to a generation of developers who thought that Software Engineering was connect the dots with a visual building blocks (maybe it will be one day).
However, in the last couple of years there’s been an emerging trend in no-code solutions on the market for B2B apps and AWS has joined the party with Honeycode. If we look at WHY we develop software as opposed to how much FUN it is to develop and put our developer egos behind us, we can see that tools like AWS Honeycode, Microsoft PowerApps and all the other no-code solution vendors make perfect sense. The cost model has changed though – you are now “renting” a platform for the life-time of that software and building on top of it as opposed to owning the software outright after custom development effort. Software engineering itself has now been disrupted with Software Platform as a Service.
At Data-Driven, we very rarely write custom user-interfaces or apps, unless specifically requested (usually for a B2C consumer app) , The majority of our insights are now surfaced in tools like PowerApps/Honeycomb, PowerBI or in a Teams App or Bot. This results in higher-quality, fit for purpose software, lower TCO and happy customers without introducing the challenges that custom apps and user-interfaces bring at this point in time.
I’ll explain why the market is shifting to this way and why Honeycomb is a good thing for everyone.
When to choose no-code over a custom app
Where no-code solutions make sense
- B2B apps (No-code general does not have a feasible pricing model for B2C or public-facing consumer apps… yet)
- Internal Apps (automate that manual scanning process you have or make an app to onboard your new employees or users etc.)
- Low number of users (which equates to a feasible ongoing cost)
- Main focus is on solving Business Requirements as opposed to a fancy UI
- Requires a quick Time to Market/Time to Insights
- PoC or Prototype where you are proving out something other than the UI
- Where the requirement is to work across all native mobile platforms and web and desktop with a single-code base.
Some real customer examples of no-code solutions we have developed:
- An AI-driven triaging process where PowerApps was a perfect fit for the UI as all the logic was in the ML models and data engineering, not the front-end
- A Teams App for integrating a legacy system with daily user workflow to reduce friction and time
- An app for BI developers to un-pause an expensive Data Warehouse on demand with auto-pausing after X amount of hours
- An Project Management tool to track project budgets and timesheets (integrated with an external system
And where no-code solutions do not currently fit:
- B2C Apps – Public consumers who use apps have higher expectations on the user-interface
- High number of users = high cost (this could be solved with alternative pricing models which I believe will come in time)
- Where a fancy UI is specifically required (usually this is where the app IS the product, as opposed to the business logic – e.g. Instagram)
For example:
- A dashboard imbedded in a public facing mobile app
- A modern web application for a SaaS product with thousands of potential users
Is no-code a new thing?
No. There’s been a few vendors in this space for a while. Microsoft attempt at Line of Business apps came back in 2010 with Microsoft Lightswitch.
We used it to build an internal project management tool combining timesheet data from a 3rd party system with project milestones to track the project for senior managers. The project was conceived and delivered into production in 2 weeks. Lightswitch itself was shut down in 2016 as Microsoft focused on their bigger strategy with the Power Platform and the release of PowerApps to replace it. The technology implementation in the Lightswitch product iteself was right at the time but like everything had a shelf-life as external market factors like technology, cloud services and advancements in web standards and browsers evolved.
It should be pointed out that there are many smaller vendors with fantastic products in this space worth checking out.
What is driving the trend behind business apps moving to no-code?
Coding is not hard. Software Engineering is.
Any good software engineer will tell you that technically coding is not hard. Most languages share similar constructs and ideas and differ mainly in syntax and maturity. Often the languages that are most used are not necessarily the best by key software engineers metrics.
Case in point: JavaScript. It allows developers (and especially less-disciplined developers) to create a spaghetti like mess of code that on the surface appears to do the job. Until you measure TCO (Total Cost of Ownership) which is not just the rapid development time. It’s the sum of the dev cost plus the extensibility cost plus the maintenance cost etc. In larger applications JavaScript becomes a mess of code that is hard to scale and slows down any new developer (did I mention that Developers are notorious for getting bored after 1 to 2 years and leaving your company due to stagnating tech? * See Side Note at bottom of post for more on this.
And yet… JavaScript is the standard web development language out there and can be used in lambda cloud functions (Azure Functions), NoSQL database functions (Cosmos DB) and mobile apps amongst other things. It’s the same thing for Python and Big Data, EDA and Machine Learning. How did that happen when we have Typescript, basically JavaScript but strongly typed!?
Because it is easier to learn. The barrier to entry for JavaScript is super low. There’s no nasty compiler errors, silly typing constraints or all the things that slow you down as a noob developer. You can focus on smashing out code and meeting the delivery requirement… at a cost that you are not usually accountable for. So the ecosystem booms and the price of labor drops. There’s a library for everything and software engineering seems to be easy again.
Which brings me to the next point. If coding is not hard… why IS Software so difficult to get right and so expensive and what on earth does this have to do with AWS Honeycomb?
Why do we have unhappy users?
Delivering high-quality software is HARD. It’s because the people with the business domain knowledge who know what the software should do and have the problem that needs to be solved are NOT the same people engineering the software. The disconnect in the middle between the business users and the software engineers is bridged by an array of agile processes, smart tooling and roles like business analysts. On the other side are smart software engineers who may or may not know the domain and who are really interested in using the latest, cutting edge technology which your company is probably not investing in due to the fast-changing nature and cost.
It is this disconnect that results in low quality software. Translation of business requirements into tech speak and back again, combined with other challenging factors like the lack of a good DevOps process equates to slow feedback cycles and the classic “Broken Telephone” syndrome where what was asked for was not delivered. Solve this problem and your TCO of Software development will massively reduce as will the happiness of your end-users and customers.
How AWS Honeycode, Microsoft PowerApps and others solve this problem
This brings us back to the problem that AWS Honeycode and others are trying to solve. The concepts is simple:
"Let the people with the business domain knowledge build the software"
Or realistically… get as close or as deep into the software engineering process as possible.
Empower your business users to build software and reduce the disconnect between the business requirements and the software delivery process. These tools are doing what Xero did for Accounting in small business. You don’t need an Accountant to do your book keeping. You need an accountant for Tax Strategy. You could still employee a Bookkeeper or do it yourself (you know your business the best) however the roles have shifted. Every role is now doing something more relevant to their skills and adding value in that particular area, which results in a better outcome for everyone.
Let's take a look at the single biggest prohibitor for the adoption of no-code solutions. Cost.
The cost of development is reduced, but at a price. This shifts to the classic “as a Service” pricing model where the cost of your software is now spread out over time.
Here’s a simple, real example from a conversion we had with a customer recently who are building a Teams App to reduce friction in their daily workflow by bridging the gap between their legacy platform and their current daily workflow in Microsoft Teams.
We started off build building a simple demo of how PowerApps can be embedded in teams as a Teams App. This took 1 day to build and demo the concept to the customer. Imaging trying to do that with React or C# in 1 day.
However, despite being a perfect match, the pricing model of Microsoft PowerApps removed it as a possible solution. Generally in PowerApps if you connect to a SQL database or custom API (which is 95% of our use-cases) then the price per user is $40 a month. In this particular scenario the customer had over 100 users of the app. Regardless of the simplicity of the app this would work out at $4000 USD per month! This is not feasible for many customers and we ended up building a Teams App using React instead, however this meant a higher development cost and slower time to market. Until Microsoft introduce a volume license for PowerApps it will be ruled out for many use-cases based on cost.
Competition is the single thing that will help no-code vendors re-evaluate their pricing models. For the same reason Slack and AWS teaming up will help improve Microsoft Teams, more competition between the cloud vendors will drive down platform prices and raise the bar on quality.
It is for this reason that we welcome AWS Honeycode to the market to keep the Microsoft PowerPlatform and other vendors innovating….
For your next line-of-business or Teams App consider a no-code solution and contact us if you need help evaluating your options and architecture. Data-Driven AI is a Microsoft Gold Partner with expert experience in data and AI and modern business applications.
Contact us to discuss your next Business Application or Teams App Integration
Tell Us About Your Project.
* Side note: Why does your company have high-turnover in the Development Team?
Why is it so hard and expensive to retain the good Software Engineers? I’ll save this for another blog post but basically software projects generally choose a tech stack and it’s meant to last a couple of years. Rewriting it in the new shiny technology that just came out is not financially feasible, so this causes a high turnover in development teams as developers leave for companies who are using the latest and greatest. The cycle of change is increasing and your average company does not have the budget to keep up with it or train their developers in it, hence the high-turn over OR high salary costs as employers throw money at their staff to compensate them for staying on old technology.
It is a disturbing trend that is increasing and causing headaches for CIOs and CTOs globally and also one of the reasons that we started Data-Driven AI. Let us stay on top of the new technology and bring the experience gained from multiple projects- it might be cheaper in the long run to scale up your development effort for a short period of time at a higher cost and then scale down again as opposed to rehiring, reskilling and trying to keep developers in your dev team. At the end of the day, good software is NOT cheap. But enough on that…
Experienced in setting up modern data platforms with advanced predictive analytic workloads. Brings strong people skills and a devops-centric, entrepreneurial approach to Enterprise software delivery.
- Hiring: Azure Software Engineer - January 12, 2023
- 10 reasons to use Azure SQL in your next analytics project - November 3, 2020
- A Developer’s Guide to Building AI Application - September 4, 2020