The 5 principles that helped scale loveholidays
Written by David Annez, Michael Jones and George Malamidis
Written by David Annez, Michael Jones and George Malamidis
Maintaining speed and innovation whilst scaling a business isn’t easy. At loveholidays this was especially challenging as our speed and innovation helped win us the award for the fastest growing business in the UK in 2019. Even with these high expectations, and against the backdrop of a global pandemic, we have not only matched our previous performance, but surpassed it across a number of important measures. Perhaps most impressively we have seen an increase in our engineers’ ability to deliver product changes by over 300%.
Underpinning these improvements were a common set of ideas which have been refined and evolved as we scaled our engineering capabilities. We have extracted these into explicit principles so they can continue to influence the changes we make as we continue to grow.
Scaling loveholidays
Scaling-up is an inflection point for any company. You have beaten the odds in finding an idea and a model that works and now have to face the new challenges that success brings. At loveholidays our success has translated into supporting 4.5 million visitors each month searching for their ideal holiday, processing over 450 billion holiday offers each day and taking a booking every 46 seconds.
Evolving a platform to efficiently and reliably handle this scale is technically complex on its own. Additionally, being part of a high growth start-up means there are other expectations and requirements to incorporate as you do. Importantly, the demand for continued innovation to maintain growth does not abate, requiring ongoing integration of new capability and experimentation with new features. To sustain this requires addressing the inevitable technical shortcuts made along the way for short term growth that compound over time to compromise the platform’s performance and flexibility. Managing this requires balancing technical and business requirements and alignment towards shared goals.
Our approach was to decompose loveholidays into smaller problems, based around our core business functions, and assemble cross-functional teams with all the required skills to solve them. This created focus and simplicity (the ability to work on one thing) for teams, enabled blending tacit business knowledge with specialist technical skills and allowed autonomy in decision making. The approach reflects the importance we place on diversity of thought in problem solving and focusing on the ends not the means. Specifically, teams with a range of experiences, backgrounds and skills will produce more innovative and better solutions. And that teams understanding the business rationale and accounting for the financial impact will make better decisions.
Implementing these changes over a 2 year period we saw a 3 fold increase in developer productivity which translated to hundreds of changes each month to our systems. To maintain the level of performance we want these changes to be consistent with an approach. However this is very difficult to manage when they are implicit and informal as at the speed we are changing it is possible to go in the wrong direction very quickly. To address this we want the principles that underpin our approach to be made explicit so that they influence our decisions and allow us to continue to improve.
Our principles
From our experience there are ways of working that produce better outcomes which have been applied to help loveholidays scale. For them to continue to make an impact over time they need to be explicit so they can be consistently applied across teams. We have done our best to summarise these, illustrate with examples and reference key influences.
Think big, deliver incrementally
We believe that you need to aim at something ambitious, sketch your path and find ways to deliver incrementally. Approaching problems this way will allow us to challenge ourselves, communicate our thinking and notice any errors and misconceptions so we can correct them along the way. Ultimately, it will lead us to create better products and systems.
Example: Our new search platform is a huge project that we’ve undertaken and it will revolutionise how we sell holidays. To better understand its impact, we delivered it iteratively, by testing it out on sub sections of our website. We did this in order to drive immediate value whilst validating the new platform’s vision.
References:
Invest in simplicity (it’s not easy)
We want to increase the speed and quality of our development by making things simpler. Simple systems are easier to communicate, understand and extend. They are also more resilient, cheaper to run and scale. Making things simple requires the ability to consider single ideas and concepts in isolation without them being intertwined with other concepts. Processes and tools alone can’t help with maintaining simplicity; it is a choice and it is one we need to invest in.
Example: We decomposed our core runtime systems into smaller, simpler ones, with ownership and support distributed across teams. We unified error reporting, making it easily accessible and actionable. As a result, incident discovery and recovery is quicker and teams spend less time dealing with production issues.
References:
Simple made easy — Rich Hickey
Out of the tar pit — Ben Mosely, Peter Marks
Leave things better than you found them
Software should be treated as a liability as well as an asset. Whilst we ensure what we build provides value, for it to continue to do so over time, there is an obligation for it to evolve. It needs to match our changing domain requirements, be updated as new technologies emerge and simplified to address any incidental complexity. Therefore, whenever we change something, we aim to leave it simpler and more aligned to our end goals than when we found. Over time these decisions will compound and help us sustain the speed and quality of development.
Example: Our evolved incident management system is an example of where we’ve ended up after focusing on improving our systems and leaving them error free. The discipline behind our incident management system and the work we do to improve systems when they fail is a core part of our culture.
References:
Focus on differentiation
We want to focus as much of our effort as possible on innovation that sets us apart from the competition. Creating strategic partnerships with vendors that share our values, using best in class commodity products and automating repeatable tasks will not only free up our time to do this, but form a platform on which we can further innovate.
Example: A/B Testing is a key activity for our business and we delivered a world first in testing cached pages “on edge” by exploring the capabilities of Fastly’s platform. Expanding our relationship with the Fastly ecosystem, we partnered with DataDome to outsource bot detection and scraping protection. Free from having to solve that problem in house, we can invest that time in USP innovation.
References:
Utilising Fastly Compute @ Edge — David Annez
Technology is a means to an end
We determine the success of our technology by measuring its contribution to the success of our business. We communicate the importance of lower level technical objectives, such as reducing latency, by explaining how they improve top level business metrics such as conversion, sales, customer satisfaction and cost.
Example: We introduced Twilio to deliver a best in class support system for our customers and replace the ageing call centre platform. Twilio enabled us to open communication 24/7 through AI, voice and chat. What marks this a successful project is the increase in customer satisfaction and the decrease in call centre costs, both of which are business vitals.
References:
Written in pencil
These principles describe our current understanding of the key elements that help you build a scalable technology organisation. They have been derived from our work over the last 2 years in successfully scaling loveholidays. However, we strive to continually learn and grow and as such these should be considered to be “written in pencil” — open to challenge, debate, suggestion and modification. As our teams apply them to a wider range of challenges we look forward to these being refined and evolved and helping us maintain our speed and innovation.
Have we missed something? Would you add to our references? Do these principles resonate? We’d love to get thoughts in the comments below
We are hiring across our engineering and product teams, you can see all of our open roles here; https://careers.loveholidays.com/