Skip to content
← Back to Blog
CloudNovember 19, 2026

Migrating to AWS with a small IT team

What we learned taking three Colombian companies to AWS: what the general manager looks at, sizing on real data and the five customer decisions.

José Antonio Jaramillo

CCO & Country Manager Ecuador / Guatemala

Migrating to AWS with a small IT team

The small and midsize companies we talk to in Colombia, Ecuador, Guatemala and the other countries where we operate tend to share the same pattern: servers that have reached the end of their useful life or are in their final stages, a pending hardware renewal, a short decision window, and a decision whose consequences will be felt not for months but for years. This usually comes with server rooms or local data centers, and with IT teams that wear several hats at once: they run operations, provide support and handle the day-to-day needs of internal users.

In recent years we have worked with three Colombian companies through this decision: a medical devices distributor and a Bogotá services company, both running SAP, and a direct-sales company operating in several countries, including Colombia and Ecuador. All three run on AWS today, and in the third quarter of 2026 they had availability of 99.95% or higher. That result also left us with several lessons along the way.

Three fronts a general manager looks at

When we talk to a general manager, the conversation is not technical. It is not about Availability Zones or instance types; it focuses on three other fronts.

What it costs. The first is economic: what the operation costs each month compared with what is paid today for hardware, virtualization licenses, support, power and space. That comparison is always necessary, and in many cases it brings to light costs that had not been identified. Two things weigh most: not buying the replacement hardware, and sizing properly, because the usual way to cover future growth is to overbuy hardware, which ends up in an inefficient use of resources.

How long operations stop. Being a small company does not mean the business doesn't run 24 hours a day or need to be up on weekends, and its operations cannot stop either. From that point of view, the agreements with the business, the schedule and the practices to minimize downtime add a lot of value, and general managers weigh them heavily.

Who runs it afterwards. How the platform is kept running over time, securely and following best practices. Internal teams are often so overloaded that they don't have the bandwidth to manage an AWS account, patch systems, check backups and monitor. That is why it is essential to deliver an operating model that gives the customer peace of mind about how their platform will be managed over time.

What we learned in three projects

Size on real data. This is a key value lever from the start, and sizing well requires a realistic baseline. The first quote is often built by matching the current servers, and that is not necessarily the best scenario, because the same hardware overbuying moves to the cloud and little is gained. The idea is to measure actual consumption, apply the vendor's best practices and size for what is really needed, today and for the future. In one of the companies, for example, the production HANA database averaged 1.3% CPU during the third quarter of 2026, with peaks of 17%, and in a couple of the cases we reviewed capacity had been overestimated.

Plan the waves well. The cutover weekend is what is most visible to the customer, suppliers and users, but behind it there is orderly, staged preparation that combines the methodology with the business requirements. In an SAP landscape, for example, we follow a specific order based on best practices, with our own methodology built on those recommendations and on years of experience: development and quality first, then file servers and interfaces, and production last.

Monitor along the way. A physical server is bought to withstand the peaks of several years. In the cloud, with the right metrics, smaller sizes can be used, and monitoring the transactional load, storage and network every month makes it possible to adjust along the way, always seeking the best balance between performance and cost. If the load is stable, a three-year Savings Plan can be committed, which is what one of the three companies did for its SAP production.

Five phases and five customer decisions

Our migration and modernization methodology has five phases, and what matters to explain is what the customer decides at the end of each one.

  1. Assessment. We measure actual consumption, inventory applications and dependencies, and review licenses. The customer receives the sizing, an estimate of the monthly cost and the wave plan, and with that decides whether to continue and with what scope.
  2. Foundation and plan. We define the AWS architecture: network, accounts, security, backups and disaster recovery. The customer approves the design and decides, based on what they want to invest, how much recovery time and how much data loss they can accept.
  3. Migration. We run the waves in the agreed order, each with testing and a rollback plan, and at the end of each wave the customer validates that what was migrated works and authorizes the next one.
  4. Modernization. Where it reduces work, we replace what the company ran by hand with AWS managed services, such as databases, backups and scheduled jobs. At the close of this phase the customer signs off and the physical servers are switched off.
  5. Optimization. Monitoring, patching, backups, cost tuning and platform support, with a monthly report. Here the internal team can focus on the business again, because monitoring and patching stay on our side, and each month, report in hand, the customer decides what to adjust.

What only AWS makes possible

There are four things a small company does not get with its current servers or with a local hosting provider.

SAP-certified instances. SAP support is maintained and sizing follows the vendor's official tables; two of the three companies run their SAP HANA production on this type of instance today.

Disaster recovery without a second data center. The direct-sales company keeps a replica of its database and application in another AWS Availability Zone, ready to take over if the primary one fails. For a small company, the alternative is usually having no recovery plan at all.

A governed account structure from day one. With AWS Control Tower and AWS Organizations, environments are separated, security policies are applied and billing is consolidated, something a server room does not offer. It is now the foundation we include in every new migration.

AWS programs for the SMB segment. AWS funds part of the migration and initial consumption for qualifying small and midsize companies. The funds exist, but a company on its own rarely knows how to access them, and as a partner we can help with that.

Where to start

What we propose as a first step is a free 45-minute assessment with the IT lead and the general manager, to review what runs today, when hardware support expires and in what order it would make sense to move things. If at the end the migration does not make sense for the company, we say so; if it does, the manager leaves with an order of magnitude to compare against the renewal quote sitting on the desk.

Book a free assessment →

José Antonio Jaramillo is CCO of Pyramid Consulting, AWS Advanced Consulting Partner and SAP Partner with operations in Colombia, Ecuador, Guatemala, Peru, Mexico and the United States.

José Antonio Jaramillo

CCO & Country Manager Ecuador / Guatemala

Más de 16 años liderando transformación SAP y Cloud en LATAM. CCO de Pyramid Consulting.

LinkedIn