Reduce Technology Costs for Your Business

As a seasoned engineer and architect, I've had the opportunity to work on dozens of corporate projects. Over the years, I've noticed a recurring problem in almost every non-trivial implementation: the domain-specific development of application logic only accounts for a small portion of the total cost, and the majority of time is spent on boilerplate code, processes, and infrastructure.
This is even more pronounced when using off-the-shelf web applications because there's little to no code development, whether it's proprietary, open-source, or custom development.

This begs the question: why do we spend so much time doing these repetitive tasks?

Some SAAS companies have done a good job at reducing the boilerplate, and have good entry pricing. But as a business matures, various frameworks or products may be appropriate at entry but fail to scale either in performance or price over time. With a smaller business, budgets are typically smaller, and freemium/cheap applications can be appealing. However, the cost benefits can drop rapidly when the business starts to grow.

Of course there are exceptions, where a SAAS is exceptionally good at a particular feature, makes it easy to integrate their product with other services, and scales pricing appropriately with the business. But many of these still hold the "enterprise" features behind a paywall too large for startups and small business.

There are also many "easy button" website hosting companies today where you can register your domain name and get a cheap monthly hosting service on a LAMP (Linux, Apache, MySql, Php) stack made popular by WordPress.
Or like GoDaddy, a hosting provider may provision you a server with CPanel and WHM installed, where you have free reign to manage the server how you like.
While there may still be use cases for this setup, I'm usually called when this solution isn't working anymore.

Performance and Security

One of the most common issues I encounter is performance. When businesses outgrow their current infrastructure, they often experience slow load times, crashes, and downtime. These issues can have a significant impact on a business's bottom line, leading to lost sales, decreased customer satisfaction, and damaged brand reputation.

In the 1990's a single server with a LAMP stack worked fine for many businesses because the internet was slow and we optimized our content to compensate, and there was little expectation of performance because everyone's site was slow.

Lyn Wildwood wrote a great article outlining some interesting performance impacts -

• Conversion rates are 3x higher for ecommerce sites that load in 1 second. (Portent)
• For lead generation, sites that load in 1 second have conversion rates of 39% (Portent)
• The average page speed of a first-page Google result is 1.65 seconds. (Backlinko)
• The average page speed of a website is 3.21 seconds. (Pingdom)
• Sites that load in 1 second have a 7% bounce rate. (Pingdom)
• 82% of consumers say slow page speeds impact their purchasing decisions. (Unbounce)
• Prefetching performed by Facebook improves page speed by as much as 25%. (Facebook)

Source - https://bloggingwizard.com/page-load-time-statistics/

To say website performance is important for your bottom-line, is an understatement.

Another challenge with relying on a hosting provider's LAMP stack is the total cost of ownership. While the monthly fee may seem reasonable, it's important to consider the long-term costs associated with managing servers. As your business grows, you may need to upgrade to more powerful servers or add additional services, which can quickly become expensive.
You'll need to continuously patch out-of-date software and configurations to avoid security vulnerabilities, and make different software "play nice" with each other.
Managing servers take a lot of time and expertise that most small businesses don't have, and paying someone to do it is like rearranging deck chairs on a sinking ship. At some point the cost outweighs the benefit and you'll need to migrate to a more efficient solution to remain competitive.

In my experience, one security breach is enough to convince most people to consider an alternative to managing legacy servers.

Another issue, is the contention of resources available on a single server.
When the database, application, and web server all share the same resources, the server itself is the bottleneck. One way to scale out would be to move the database onto a new server so it's not fighting for resources. This was a typical process I'd done for apps prior to ~2015 (a new approach is described further down).
But now we have 2 servers that need patching, security scans, etc... and the total cost of ownership is growing fast.
Assuming your business continues to grow, eventually we'll need to consider how to further scale a database beyond a single server.

I've not even touched on server failover or high availability, which adds more complexity and cost.
Suppose you have high traffic only during a certain time of the day.
Do you scale everything up so that your system doesn't crash during peak time? How do you feel about paying for resources all the time, that are only used 10% of the time?

Imagining a freeway is a good example. During "rush hour", all the lanes get jammed up, but they may be mostly empty the rest of the hours of the day.
As drivers, we like moving down the road at a steady pace.
But as a business owner, if those lanes represent CPU and Memory resources, we don't like to pay for resources that aren't used.
What if the freeway lanes could replicate themselves depending on the amount of traffic? So that you always get somewhere at the same rate no matter how bad the traffic gets.

A modern approach

Dynamic horizontal scalability may not be possible with freeways yet (darn), but modern technology like "containerization" and automation make this possible with our web applications (nice!).

In contrast to legacy server installations, containerized workloads have a thin operating system "baked in" to the application image, or in some cases no operating system at all. Using hardened security base images also makes it easier to prevent vulnerabilities.

I've built dynamic scaling for many companies, including scaling custom apps and various web processes and applications.
In one case, for a rapidly growing company, I moved several WordPress sites off GoDaddy and into this dynamic architecture. We found that we needed 5x the amount of resources around lunch and dinner time, and the rest of the time there was minimal traffic.
It would take another large article to describe the case study, but in summary, we built a modern data lake/pipeline/and warehouse, Kafka message bus, deployed several WordPress sites, a custom application for managing their business data, an internal "Enterprise Portal", a digital asset manager, a full oauth solution for SSO, and several other apps and event driven processes. The total monthly infrastructure costs are around $2.5k. This is a fraction of what it would cost to piece together different SAAS products.

As an example - WP Engine offers an enterprise level package that costs over $66k/year.
https://aws.amazon.com/marketplace/pp/prodview-olkg3x54zujdc

That's just WordPress. If you want enterprise grade SSO, an internal portal, or other apps, you'll pay significantly more - and need the expertise to fit them together.

We've done the hard work

Considering the cost of enterprise grade services, we built a custom platform on the cloud with cost, scalability, and security as primary considerations. We expect to grow, and our IT costs should not outpace our profit!
While it required a significant investment of time and resources upfront, the long-term benefits have been well worth it. By managing our own infrastructure with modern IAC/containerization/automation, we have complete control over the technology stack, ensuring that it's tailored to the specific needs of our business. Plus, we've been able to optimize performance and reduce costs by eliminating unnecessary services and processes.

Of course, building your own infrastructure isn't for everyone. It requires a certain level of technical expertise and a willingness to invest time and resources upfront.
This is why we're now offering our platform as a managed service. Using modern automation, we can onboard a company with enterprise grade services for an entry price less than a typical enterprise plan, for any single product.

Features like "Single Sign On", failover servers, load balancing, dynamic scaling, and engineering support are important to all business - not just large enterprises. Our pricing model is based on resource utilization, rather than a predefined set of resource allocations that most providers offer.

We're democratizing enterprise grade services for small to medium size businesses. At Savantly, we believe small to medium sized business should have enterprise grade services without the enterprise price tag.

Contact us to get early access to the platform and help us shape the product with your needs. We're currently accepting partner applications.

Ultimately, every business is different, and there's no one-size-fits-all solution. However, by taking a thoughtful and strategic approach to technology management, you can position your business for long-term success.

Apply for early access