Stay is a Guest Experience OS that provides access to all hotel services from guests’ smartphones. Founded in 2012, the company serves more than 1,200 hotels in 68 countries and continues to expand.
With solid financial health and consistent profitability, Stay has operated as a bootstrapped company since its inception. Its product is deep and inherently complex, offering a comprehensive range of services and capabilities for both hotel guests and staff.

Over the years, the company has undergone significant technological evolution, leading to decisions that have increased the complexity of its tech stack. This continuous innovation has been essential in maintaining Stay’s competitive edge in the hospitality tech industry.

As my time at Stay wraps up, I can’t help reflecting on an incredible year.
In September 2023, after several conversations with Joan (CEO), I had the opportunity to join Stay as a Tech Advisor. At that time, the company was seeking leadership to tackle its various challenges, especially around scalability and establishing robust technical processes.

Despite the magnitude of these challenges, it quickly became clear that a part-time role wouldn’t be enough. As a result, in December 2023, I joined Stay as full-time CTO, fully committed to leading the company through its next phase of growth and technical innovation.

Initial diagnosis

In my first month, I focused on adapting to the company and gaining a clear understanding of its main challenges, including productivity, scalability, knowledge transfer, and team design.

After identifying and outlining the key issues, I presented a strategic plan to the C-level team, focusing on productivity, infrastructure, security, reducing tech debt, and streamlining operations, with clear milestones and timelines to achieve these objectives.

Productivity

Early on, one of the main priorities was stepping up productivity and cutting down on friction. Given the complexity of our tech stack, shaped by the depth of our product, there was plenty of room to work on improvements.

One key enhancement was consolidating all micro-repositories into a single monorepo, organizing each project into separate folders. This significantly reduced maintenance overhead compared to managing 20 individual repositories.

We cut down on branch maintenance complexity by moving away from git flow in each micro repo and switching over to git trunk based development in the main monorepo.
At the same time, we took the chance to switch out Bitbucket for GitHub, reworking all CI/CD pipelines to fit into the new setup.

Moreover, we restructured our pre-production infrastructure, moving from three outdated staging environments to dynamic, team-specific setups defined through Infrastructure as Code (IaC). This enhancement ensured scalable, replicable environments that closely mirrored production.

Infrastructure

We identified significant areas for improvement in our infrastructure. Despite being hosted on AWS, management was largely manual, lacking a clear track record or processes for upgrading critical services or recreating staging environments that mirrored production.

With the company setting ambitious goals, such as achieving compliance with security standards, having exposed services like RDS accessible over the internet was a significant roadblock. To address these issues, we created a detailed roadmap outlining the necessary improvements and partnered with a new provider that had the resources and expertise to implement these changes efficiently.

We decided to start a new infrastructure in parallel, ensuring the old setup continued serving clients without interruptions. This approach could have significantly increased our AWS costs, but with an AWS credit grant from Bonsai Partners, we mitigated the financial impact and executed the migration seamlessly.

Key milestones achieved:

• Rebuilding the infrastructure under a new AWS account structure with separate accounts for Audit, Staging, Production, and root management.
• Implementing Infrastructure as Code (IaC), ensuring all changes are stored in a repository, with no manual changes in the AWS console.
• Creating staging environments identical to production, scalable as needed.
• Reducing costs by using spot instances in staging.
• Enabling Google Sign-In for AWS access with token expiration, adhering to a zero-trust model.
• Implementing Twingate as a VPN solution, granting resource access based on team roles.
• Establishing monitoring using the Grafana ecosystem, including Loki and Prometheus.
• Upgrading the main MySQL database to V8 (in RDS), MongoDB, and EKS clusters, while implementing a future-proof upgrade strategy to minimize service disruptions.

CDN migration to Cloudflare

While some services were already delivered through a CDN, clients frequently reported issues with performance and reliability. Additionally, the previous CDN offered limited security features, which failed to meet our evolving needs.

We migrated to Cloudflare, enabling all services, including APIs, to be served through their robust platform. This transition not only enhanced performance and availability but also leveraged Cloudflare’s WAF, providing a significant upgrade in security capabilities.

Security

Despite the critical importance of security for Stay’s clients and the need to keep up with new client requirements for audits, we stepped up our efforts by significantly increasing the budget on cybersecurity to ensure compliance and readiness.

Key initiatives:

• Preparing the entire company for SOC 2 Type II compliance by defining processes and policies.
• Negotiating with Vanta to secure the platform at a highly competitive rate.
• Establishing a partnership with an auditor for annual recurring audits.
• Enhancing observability through logs and OpenTelemetry.
• Deploying SentinelOne for endpoint detection and response (EDR).
• Using NinjaOne to monitor and manage all company workstations.

Reducing Tech Debt

A strategy was drawn up to cut down on tech debt caused by microservices complexity and to phase out redundant stacks like Memcached, Redis, SQS, Kafka, and others. The challenge was carrying out this plan while keeping up with the product’s ongoing development and evolution. To speed things up, a dedicated team was put together, freeing up other product teams to focus on their main responsibilities.

Operations

The operations strategy focused on designing and structuring new teams to ensure effective knowledge distribution across the organization.

To maintain continuous service availability, a 24/7 coverage model was established, with the infrastructure company handling first-line support and escalating issues to Stay’s engineering team when necessary.

Closing

While the company has been around for 12 years, I only got to be part of it for one, yet I always felt fully included in the journey.

I’m truly grateful for the opportunity to pitch in, even if just for a short while. A special thanks to Javi (VP Eng.), A. Valera (EM), and the entire product and engineering team for backing me up every step of the way. Stay is the kind of place where you can easily reach out to the founders and the rest of the C-level team to talk things through. It’s been an unforgettable experience, thank you for giving me the opportunity to be part of this incredible journey!

For those who are unfamiliar, Exoticca is a leading tech travel company in the multi-day tour package space. We’ve experienced hyper-growth, operating mainly as a B2C business in 7 markets, with a targeted revenue of 200M in 2023 and a positive EBITDA. However, the journey to reach this point has been quite challenging, with many bumps along the way.

When I joined, the company was about 70 people overall, with just 6 in technology, infinite roadmaps, and amazing ideas but clearly a lack of alignment. Lucky me was in a winning horse with senior Executive members and enough budget for me in technology to start building teams.

Currently, our technology department is about 65 members, who are divided into 8 teams.
Each team is carefully balanced to guarantee being autonomous and deliver value. The areas of expertise each of them have are mainly: Backends, Frontends, DevOps, QA, Data Engineering, Data Analysis and Product Owners.

The beginnings

My first homework was to design a structure to fit all the company’s projects and start delivering value as soon as possible. Back in that day’s technology department was responsible for all, from the website, to backoffice, internet connectivity, and even the printers! :)
Although there was no clear alignment between Product, Marketing, Sales, and Tech, the company’s finances were growing and consistently meeting the budget. However, as the headcount of the company increased, the scalability of its processes began to suffer.

We split out into teams based on the domains Exoticca has: Web, Backoffice, Flights, and IT Support. In the next iterations, we added extra teams such as Data, and Travel Experience.

The composition of the team was only possible because there was a clear roadmap in place and specific objectives for each team member to understand and work towards. In my previous experiences, I have worked with OKRs as a way of achieving these objectives, either because it was already the established framework or because I helped the company adopt it. In my current role, I took the opportunity to educate the leadership team on how to implement OKRs throughout the organization. Since then, it has become the standard framework we use to guide our work towards meeting our objectives.

In addition, I realized that my department had no knowledge about SCRUM. Being a fresh SCRUM Master certification holder, I introduced the methodology to the team. I have to say they quickly embraced the benefits of SCRUM, such as promoting collaboration between technology members and business stakeholders. One of the most expected meetings for all the company happens every two weeks by the Sprint Review.

The culture of a CTO

Despite the common responsibilities that we typically associate with a CTO, such as driving innovation, aligning technology with business goals, and developing and managing teams, I also came to the realization that it’s important to have knowledge in areas that may not fall directly under my purview. For instance, issues with office networking and internet connectivity can have a significant impact on our business, particularly since we deal with VoIP.

I encouraged the team to help each other out, regardless of their specific areas of expertise or skills. Working with a team that had strong values and accountability made it easy for me to spread this message to new members. While it can be easier to instill this kind of culture in a smaller team, it’s important to continue to foster this mindset as the team grows.

Milestones

As we reflect on the last four years, it’s clear that we have achieved some truly outstanding accomplishments. Here are just some of the highlights:

• We relocated our company’s headquarters to a new location and built the entire network infrastructure from scratch. The new office had multiple levels, which added complexity to connecting the different floors.
• We migrated our website from a legacy stack, which used jQuery and outdated basecode, to a modern React-based stack, leveraging Next.js and server-side rendering.
• We leverage IaC tools such as Terraform and Ansible to write our entire AWS infrastructure as code, thereby ensuring that there is no need for any manual intervention through the AWS GUI.
• We implemented a robust CI/CD pipeline that automates the build and deployment process across the cluster, allowing any technology user to deploy updates with just one click.
• We developed one of our company’s core features, called Flights Hub, which facilitates connectivity to various GDS platforms, including Amadeus and Travelfusion. This integration provides us with a vast flight portfolio and more competitive pricing options.
• We upgraded our codebase from PHP 5.6 to 7.4, and later to version 8, resulting in a significant improvement in the performance of our backend applications.
• To improve our monitoring capabilities, we centralized all of our log management into Datadog. This allowed us to have a single view for deep diving and made it easy to set up alerts for notifications.
• We have established a strong culture of monitoring and observability, utilizing tools such as Datadog for log management and Prometheus for health metrics. We consult this data through Grafana and regularly monitor the uptime of all our services.
• During the COVID lockdown, we took advantage of the reduced website traffic to migrate from AWS Frankfurt region to N. Virginia. By doing so, we were able to bring our services closer to our primary markets in the US and Canada.
• To enhance our organic search, we consolidated from having multiple market with specific domains into a single .com domain that improves the SEO reputation.
• We designed the website architecture being modular, allowing for strategies such as white labeling without the need to reimplement the code.
• We ensure high availability for all our websites by using auto-scaling groups that are tailored for each market. We rely on Varnish and Redis mainly.
• We invested in improving QA by implementing E2E testing using Cypress and Datadog Synthetics.
• We established a Data team from scratch and implemented a Data warehouse along with pipelines to fetch data from various sources.
• We developed the native in-destination application from scratch, starting with iOS and eventually expanding to Android.
• At the same time as expanding our teams, we established new tech lanes and roles such as Tech Leads and Technical Directors to provide middle management and alignment.
• We allocated a budget for training and development for each technology employee to support their career growth. Moreover, we sponsored various technology events in Barcelona, including DevOps Barcelona and PHP Barcelona.
• Build remote development, and staging environments are fully functional, using the same architecture and software as the production environment.
• Real-time data-driven decisions enable us to take immediate actions that have a direct impact on the user experience, such as adjusting prices in response to fluctuations in the market.

Down-to-earth architectures

At Exoticca, we primarily use three programming languages, namely PHP, JavaScript, and Python, for our tech stack. Furthermore, we have established a seamless architecture across all teams to ensure that everyone can comprehend how services from other teams function.

We were tempted to explore the use of other languages such as Go or Rust, especially for new services. However, we were cautious about adding unnecessary complexity and instead focused on ensuring that we had enough skilled personnel to manage the current tech stack. As a result, our teams have strong expertise in JavaScript for both frontend with React and backend with Node.js, as well as in PHP and Python.

To ensure both high productivity and availability of our application, we prioritized implementing the hexagonal architecture in our primary services. Additionally, we placed emphasis on creating seamless AWS infrastructure architectures that were familiar to all technology employees.

In summary, by having a concise technology stack, utilizing hexagonal architecture, and maintaining consistent infrastructure across all teams and services, we enable seamless transitions for employees moving between teams. For example, someone from the Web team can easily transition to a new team and quickly adapt to new product challenges, without experiencing significant friction due to differences in technology and infrastructure.

How I maintained a work-life balance during the past four years

When I joined Exoticca, my first daughter was just four months old, and it was a challenge to navigate being a new parent while also starting a new job. I vividly recall the daily struggles and exhaustion from lack of sleep, but as they say, one child is not a party, right? In June 2020, in the middle of the pandemic, trying to recover a travel company, my second daughter was born.

During my paternity leave, I remained connected to my responsibilities in the company, primarily due to the challenging circumstances that the company and the sector were experiencing at the time. As a result, I struggled to maintain a healthy work-life balance. However, looking back, I must say that it was quite fun!

And finally, I just want to give a huge shoutout to everyone who has had my back during this whole journey — you guys are awesome! I’m a bit worried that I might have missed someone, so please don’t be mad at me.

Jesús, your remarkable vision and passion have led to the creation of an exceptional company, and thanks for trusting me. Pere, I appreciate your unwavering trust in me from the very beginning, as well as your support and resources that have enabled me to succeed.

Lluís, your broad perspective and ability to handle all aspects of the company are truly impressive. Laith, you have revolutionized our MKT strategy despite the challenges posed by our complex product. Mari Carmen, your ability to navigate the roller coaster at the same time building a solid HR foundations.

Joaquim, your support as my right-hand man has been invaluable since day one. Carlos, your transition from development to product owner has given you incredible abilities. Matej, this was our second journey together, is always a pleasure to work with you. Raul, your rapid growth in technical skills and people management is impressive, and I encourage you to keep up the good work. Marlon, you have an exceptional ability to integrate additional components into the complex system you manage. Alejandro, because the office infrastructure matters a lot. David, you bring a data-driven culture and enable the company to work on that way. Isaac, managing a complex domain like flights is not easy, and you have done an outstanding job. Ricard, I’m deeply grateful for the invaluable support you provided during the initial stages of our infrastructure development.

The list goes on, and I am grateful to each and every person who has helped me along the way.

In this article, I will explain a practical case in which OKRs were implemented in a company where most of its employees were not familiar with this framework, and used their own resources to achieve their goals. This draws from a years’ experience using OKRs.

I will start briefly explaining what OKRs are — it will be succinct as there is a lot of information available and my aim is to explain how to implement them in a real case scenario.

O — Objectives
K — Key
R — Results

According to Wikipedia: “Objectives and key results (OKR) is a framework for defining and tracking objectives and their outcomes”.

John Doerr popularised this term when he transferred his knowledge from Intel to Google, where it became popular, and other companies such as LinkedIn, Zynga, Twitter and Uber have been using it ever since.

As mentioned, my goal is not to explain what OKRs are –there are many articles and books for this purpose– but to show you how to implement them in an organisation. However, I’d like to recommend the following resources as further reading on the topic and the concepts involved:

Books

Measure What Matters: OKRs: The Simple Idea that Drives 10x Growth
Author: John Doerr

Radical Focus: Achieving Your Most Important Goals with Objectives and Key Results
Author: Christina Wodtke

Articles

The Art of the OKR

The Guide to Company Objectives and Key Results (OKRs)

Videos

Startup Lab workshop: How Google sets goals: OKRs

What is Social and Beyond? About this company

“Social and Beyond brings technology together to turn the Wi-Fi network into a powerful marketing tool by creating a link between free-access Wi-Fi networks and their users.”

I joined the team in April 2017 as a CTO with the aim to strengthen the technical team to help deliver projects that were in progress, but my role extended to other functions, as you will see.

I found that, despite being a small company of about 25 employees, the communication between departments was very limited and poor. Developments were ordered by a particular team and, once delivered to production, it turned out that they were not needed. A lot of time and effort were devoted to tasks that did not bring any value to the client or end user.

All in all, each team was heading in its own direction and without a clear goal uniting them.

My previous job was at Trovit, where I learned a huge amount and it was precisely there where I found out about OKRs and, honestly, once I became accustomed to it, working in an organised manner using any another way felt less successful. Trovit had about 100 employees back then and each team set its OKRs each quarter, while these were aligned with the company’s overall goals.

As a tech lead at Trovit, it was very easy to use OKRs, as I only had to focus on designing them for my team — much work was already done!

However, at Social and Beyond, the challenge was enormous. We did not have a company mission or vision, all teams presumed they were the most important one and nothing was followed up. The technology team had completely different goals every week and people were very under-motivated.

It was clear that it could be a great opportunity to help implement what I had learned at Trovit through OKRs.

Preparation and implementation

It was obvious that the way in which we were working was sub-optimal and there was much room for improvement. It was not about the team itself or how good they were at their work — instead, we were losing out due to poor organisation. It is often said that when someone comes from outside, without a contaminated vision, it is easy for them to identify these shortcomings.

It was evident to me that there was no way that could work — as I was used to working with OKRs, it seemed obvious they should be implemented here too, but the challenge was to transfer them to the company and that everyone would be willing to embrace them.

The first thing to do was talk to the board of directors about it to so that they could understand the need for a better organisation, which had to be interdepartmental, and that everyone comprehended that we had to work together. The key to OKRs is not only that it guides you and gives you focus towards a particular goal, but it does so with total transparency where any member of the team knows and understands the goals of all other teams.

To do this, the most helpful thing in these instances and possibly in all cases, is the data. To be able to check with numbers where we stand in terms of the delivery of a product and the time spent, and to compare it with previous deliveries, gives a clear vision of where we are today and the extent of the challenges faced.

The company always wants to improve and obtain a greater yield, so it’s relatively easy to show that we need to implement a methodology or framework that helps us in this regard — especially when underperformance is brought to the fore.

More lines could be devoted to the topic of working with a vision and mission — as you know, there are a lot of articles and books and I wouldn’t want to bore you with this. I would only like to emphasise that it is very important to set a mission as we will build upon it to work with OKRs.

The last thing to do now is to explain to the board of directors what OKRs are and how to apply them. The key is to have done all the work beforehand and anticipate all kinds of questions such as: “Why can we not work on more goals?” or “What happens if I want to change the goal in the middle of the quarter?”

Everyone can take part and their ideas are key to setting the goals

Before any endeavour to explain OKRs to the team, it is a good practice to have them participate in the definition of the future goals and to do so, there is nothing better than working sessions where they give ideas or suggest improvements for our products.

This is perhaps one of the most magical moments, where you realise that everyone is keen to suggest improvements — when they are offered the opportunity to do so, you really see the team’s potential and you realise that it is they themselves that make the company grow. People are the most valuable asset!

How to set OKRs

As it is a framework, it must help us to outline the rules that everyone can understand and follow.

• Company goals are shared with everyone
• OKRs are effective quarterly
• Each team will have a maximum of 3 goals each quarter
• Each goal will have a maximum of 3 key results
• It will be monitored monthly and goals that are met will be presented to all
• We must meet a minimum of 70% of each goal

Goals must be aspirational — the team must work hard on this aspect and make each goal their own while seeking full commitment to achieve it.

The key results linked to each goal must be measurable or quantitative. They can be of a binary type –achieved or not achieved– or rated 0 to 100, for example, where we can measure performance every month.

We can measure them based on growth, engagement, revenue, performance, and quality.

Setting up the teams

Teams must be self-sufficient and large enough to be able to achieve their OKRs.

One of the main problems of any organisation, large or small, is interlinked goals that are dependent on each other goals, which often exist and must be seen as a challenge when it comes to achieving company goals.

In small companies such as Social and Beyond (25 employees), communication and the resolution of conflicting goals should be straightforward. To help with this aspect, each team will have a “lead” figure. The “lead” will be the person who will talk with the other teams to help resolve issues and prioritise work to achieve the goals.

Once a week, a meeting will be held with all team leads taking part.

Each team sets each quarter’s OKRs based on company goals, so the team is responsible for them and engages with what needs to be done in the following 3 months.

Each team can work as they wish in order to achieve their OKRs. For example, at the technology team, we applied SCRUM with our Jira dashboard, but other teams did it with Trello or even some used spreadsheets. OKRs serve as guidance and point at what we need to focus on during those 3 months, but the particular actions and how the team meets the goals are up to the team itself.

Marking all events on the calendar

The forthcoming days and months are marked in the calendar for the whole company — in this way we can forecast appropriately the following events:

• A week devoted to the setting of the current quarter’s team OKRs
• The last day of each month in which goals that are met will be presented to all
• Global assessment at the end of the quarter in which challenges for the following quarter will be explained
• A week devoted to the setting of team OKRs for the following quarter
• Weekly meeting with each “lead”

Tools

As always, less is more, and in this regard, the simpler the tool we use to stay focused on goals, the better for us.

We take advantage of GSuite so that we all have access to the spreadsheet, which allows us to share it with everyone and make it editable by the “leads” of each team — great!

A spreadsheet available to the whole company is created, which should include as many tabs as the teams we have. Each team must set their OKRs so that everyone can see what team A or team B are doing at all times.

Here you can see the template we use, which consists of 7 columns: Objectives, Key Results, Weight, Accomplishment, Month 1, Month 2, Month 3.

The “Weight” column allows us to set the most important goals for different reasons, for instance: the client expects a delivery during the first month of the quarter, or it has priority so that other goals can be met.

A follow-up is done each month and its value is logged in the corresponding column for the current month. Then this will be automatically calculated in “Accomplishment”.

Feedback

After a full year –therefore four completed quarters using OKRs within the company–, the team results and feedback are very positive.

Obviously, the first quarter is the trickiest, as everybody must make an effort to understand and be willing to apply them because it is a cultural change. Improvements are visible from the second quarter onwards.

One of the key outcomes after using this framework is work focus. Everyone in the company knows where their efforts are placed and there is no change of direction. This allows teams to become organised and independent, and therefore they are committed to the goals and the company.

We have learned to say NO to certain projects to be able to focus on those to which we have committed. To recognise that you cannot have it all is perhaps the most difficult exercise for a company, but reality is always more challenging than we would like. OKRs allow for transparency and visibility among teams, so it also helps to see the capacity of a team for delivery, and based on that, we will have to rule out projects simply because there is no time to do them. When the decision to say NO is made, then the company understands that we must value that we are aware of what our company does best, and what customers and users will use or buy.