Established corporates often have difficulties adjusting their business efficiency to fast changing markets. In these dynamic environments start-ups are often the ones that excel. A discussion of four ways large corporates in fast changing environments can learn from the successes of start-up companies.
Organisations grow by finding new markets, customer segments and products, by diversification of activities or through autonomous growth. Every mature company needs processes, management layers, roles and responsibilities that help to keep it on the right track. Until recently these companies, despite their size, did not necessarily struggle to adjust to their gradually changing surroundings. With a focus on solid performance and the continuation of the company, corporates experienced steady growth despite their complex organisation.
However, due to the rise of (mobile) internet and fast digitalisation, major changes have taken place in a large variety of markets, e.g. telecommunications, music industry, retail, media, publishing and financial industries. Especially in these industries, corporates with a traditional focus on efficiency have difficulties adjusting to their environment.
Large corporates challenged by ever-changing environments often exhibit familiar symptoms. Their complexity slows down the decision making process, which eventually completely freezes due to a fear of risks. An increasing number of stakeholders in the matrix-organisation and the formation of stand-alone department or business units cause an increase in complications. Larger companies have relatively less employees in direct contact with the customer. Moreover, decisions are still too often based on intuition instead of data.
Blank & Dorf (2012) define a start-up as an organisation intending to find a repeatable and sizeable business model. Start-ups do not face the problems mentioned above, which enables them to withstand the dynamic environment relatively well and allows them catalyse change. In the telecom market, start-up company Whatsapp is thriving at the expense of major telecom providers. The sending of text messages, once one of the cash cows of the telecoms, is practically made redundant by the rise of the messenger app that can be found on most smart phones. Successful start-ups continuously improve themselves by taking pre-calculated risks, by having engaged employees working in multi-disciplinary teams, by prioritising based on data, and by maintaining daily contact with the end-customer. Major corporates could greatly benefit if they learn from the successes of start-ups all over the globe.
Large corporates tend to take less risk, as damage of failure is often high. Managers map new business plans in great detail and sketch different scenarios before taking action, often in business settings isolated from the real world. Plans cover long periods of time and projects are often realised through waterfall-like methods. Many unknowns are predetermined through assumptions, even before a product is manufactured or a new idea is implemented. The belief prevails that defining an idea or product completely on beforehand increases the chance of success.
This risk-avoiding long-term planning strategy worked in relatively static environments that were predictable to some degree. However, it is no longer viable in today’s reality. Indeed, sticking to a pre-defined plan can be very risky in a changing world. Adaptation is the key to survival. Flexibility requires an experimental way of working that occasionally includes smart risk taking. This is exactly what start-ups do: it is in their nature to continuously reinvent themselves. As the product-market is still unknown to many start-ups, they need to develop a framework to find the product that meets customer demands. Taking calculated risks sometimes results in success, but just as often in failure and causes many hick-ups along the way. This is not necessarily a bad thing: as long as the organisation learns, it is a step closer to a good product-market fit.
Most start-ups use a specific hypothesis-based method to find market demands and develop their product. Instead of developing a complete product, a small but vital part of the product is quickly developed. Through testing whether this minimalistic acceptable version of the product meets market demands, risks are limited. This principle is called MVP: minimum viable product.
A MVP can be used to test if certain presumptions about customer needs are justified. It is therefore important that a MVP is a minimal but working version of the final product that the customer desires. Eric Ries (2011) distinguishes several kinds of MVPs in The Lean Startup. Within the first kind (concierge MVP) the customer experiences a completely normally functioning product, however no investments are yet made in automating and elaborating the processes. At the front side the customer is given the impression of an automated process, but at the back everything is still executed manually by one of the company’s employees. Another example is the waiting list MVP. A button on a website allows customers with a certain need to subscribe themselves to a newsletter. Only when there is enough interest for the newsletter, a new product will be developed. In the meantime, customers can indicate their wishes for the final product. These methods use a minimal amount of resources and time to develop useful insights about the product and targeted customers. New hypotheses are formed which will again lead to improved versions that will be re-tested. Designing and developing products like this is flexible and interactive. One cycle only takes a few days to several weeks. Interim results are quickly visible. Invalidated hypotheses can be adjusted prior to finalising results. Start-ups use these methods to quickly and efficiently develop new ideas and concepts.
For large companies, working with this new method requires a change of mindset. The main focus is no longer to do everything within a certain budget or time (e.g. implement project X within Y months), but to do it well enough as quickly as possible and from there on improve further. Hypotheses are continuously formed and validated (or not). If something works, it can be scaled up. If it does not work, the company can still look for other solutions. The launch of a product is no longer the finish line, but the start of the next iteration of building, measuring, and learning.
To successfully implement this new technique, employees should be given the freedom to execute data-based experiments. Processes that enable quick iterations, like mechanisms that enable fast feedback-loops and integrated quality assurance are extremely useful. Employees working on the product must be able to test and validate quickly. Employees with direct customer contact must be able to carry out experiments directly without the interference of for example the IT-department. The possibility of easy rollbacks is required in case something goes wrong.
Booking.com, the market leader for online hotel bookings, is renowned for its continuous improvement processes. In 2010, Arthur Kosten, the company’s marketing manager at the time, named the most important reasons for the success of the start-up. Decision making is not based upon HIPPO (Highest Paid Person Opinion) but upon the actual actions of customers on their website. Final alterations to the platform are only implemented when a hypothesis has been validated through data. Booking.com often launches a new version of its website. Moreover, new versions are always tested against the original to learn how customer behaviour changes as a result of the alterations (Kranenburg, 2012). Booking.com’s search box on their homepage has evolved over the past years by hundreds of iterations, resulting in better searches and higher conversion rates to bookings. It’s interesting to compare the frequency of releases at Booking.com (more than a thousand each year) with that of operating system Windows (one major update each year). The removal of the start button in Windows 8 was heavily criticized. Due to the low frequency of releases it took over a year before this customer feedback was implemented and the start button to make its comeback. At Booking.com this would have never happened.
As companies grow they form separate departments, business units and internal advisory bodies, it becomes more complicated to enable clear communication between departments, to have everybody looking in the same direction and to utilize the knowledge and expertise of all disciplines. All of this is a lot easier in organisations with a well-defined team size. All functions work together in close proximity, making everyone aware of each other’s activities. A change initiated by one team member with a certain function is easily picked up by others with different functions. Feedback loops are short so necessary improvements can directly be taken care of.
Successful start-up teams focus in their collaboration on improving the one metric that matters (Croll & Yoskovitz, 2013). This is the leading metric for predicting success in that team, and depends on the type of company, the business model and the development phase of the company. Fact is that everyone in the start-up works in service of the improvement of that single metric by delivering knowledge from his or her own area of expertise.
Working like this is a bigger challenge for larger corporates. Department goals can vary considerably or even collide and teams are usually physically apart from each other. For top-management it is harder to get a clear overview and therefore, it would be better if teams looked for solutions autonomously. Managers should promote cross-divisional collaboration by putting together (temporary) multi-disciplinary teams. Just like in a start-up, these teams then focus on solving a single problem or improving a single metric. The customer’s best interest is key in this. It is of importance that multiple disciplines are involved in both the formulation of the solution as well as its implementation. This allows problems to be discussed from multiple perspectives. Possible barriers to implementation are noticed early on. When a set of related problems is solved, a team can start on the next challenge, possibly in a new context or composition. On top of the team-based work, employees with comparable functions should meet regularly to exchange essential knowledge and experiences.
Spotify is a music streaming service with forty million active users of which a fourth pays for the service. Since its establishment in 2006, Spotify grew from thirty employees to over 1200 in 2014. This growth required a controllable organisational structure that enabled employees to align quickly and to keep on innovating (Kniberg & Ivarsson, 2012). Spotify’s answer was as musical as their product: they based their organisation on the interplay of a jazz band. Jazz bands exist of autonomous musicians (the teams) that simultaneously play in order to form the eventual song (the total Spotify product). Local creative outbursts are allowed, desired even, as long as it contributes to the whole. Therefore, Spotify’s fifty teams are encouraged to act as autonomous as possible but to always stay in line with the strategy of the overall organisation.
Spotify makes sure that every team sees themselves as an autonomous start-up: a team consists of five to eight persons, encompasses all functions to operate independently and has their own office space. Multiple teams together form a tribe that consists of up to one hundred people. All teams within one tribe are located in the same building but in separate offices – the role of a tribe is similar to that of an incubator at start-ups. To mimic the social and informal element of working at a start-up, Spotify has guilds: tribe-transcending interest groups with topics such as leadership or online security. These guilds facilitate mainly informal contact between the teams and tribes through e-mail lists and by organising get-togethers and small conferences. Any interested employee can join a guild. This decentralized structure not only has a positive effect on the flexibility of the organisation. The ‘start-up feeling’ also contributes to the motivation of the employees. Last year, a survey showed that 94% of Spotify’s employees are satisfied: there is a reason for Spotify being known as one of the best employers among start-ups in the technical industry.
Prioritising ideas based on data
In the past, companies usually knew little about the market and their customers. Recently the amount of data available for companies is increasing, but many still fail to correctly aggregate this data. The surplus of data and ideas causes some managers to not be able to see the wood for the trees anymore. In the worst case this causes one to completely freeze: one does no longer take any decisions or chooses to compromise.
A solution to this problem is not to directly compare all ideas, but to prioritise the problems first. Successful start-ups make use of data to determine the importance of each problem. They realise that usually a small part of the problems has the largest impact. For example, within a service organisation several issues with one product can cause the majority of complaints. By using the correct data, the importance of each issue can be clarified in a Pareto-curve.
Many managers in companies skip an important step: they do not validate the problem. People sometimes forget that a single observation does not necessarily indicate a systematic problem. To validate a problem, managers need to formulate a hypothesis from the perspective of the customer and then substantiate it with patterns in data consisting of customer behaviour, customer feedback, or additional anecdotal evidence. Only when support for the problem hypothesis has been found, one can speak of a validated problem that can be solved. Successful start-ups do not only hypothesize and validate the problem but also the solution. Possible solutions to the problem are formulated as hypothesis. For every hypothesis the ease of implementation and the expected improvement potential are estimated. The combination of these elements determines which solution obtains the highest priority.
To validate a solution customer behaviour data is needed. In the current digital era, one can collect this data in several ways, e.g. by running different versions of a website parallel to each other (A/B or multivariate testing). Another method is to build a prototype that is just good enough to learn as much as possible about customer demands (the aforementioned MVP). In this case, ‘less is more’ holds; when in doubt about adding a certain function, omitting it is almost always the best choice. After all, one can always add it in the next improvement cycle. Using user- and customer surveys and online analytics, companies can easily test which ideas will have the most impact, before they are fully implemented. This also decreases the operational pressure, as not every idea needs to be elucidated to determine its effect. Even with a small amount of data retrieved through these methods, decision-making processes of managers can be greatly improved.
Etsy is an online market place for hand-made products. Interested parties can buy art, accessories, jewellery and vintage clothing online, directly from the producers and artists. The company uses a method of predictive modelling to estimate the remaining customer lifetime value (CLV) of every customer, as described in a case study by Medri (2012). All departments contribute to improving the CLV. By making a solid estimation of expected future spend, they can determine how much resources should be allocated to retain this customer. New ideas are first prioritised according their contribution to the CLV. By doing this, Etsy identified ‘golden’; customers with a higher CLV and rewarded these customers with extra service. Another campaign was specifically designed to activate customers who hadn’t bought anything during the past 60 days (thus with a low expected CLV). The relatively simple campaign delivered them a direct additional benefit 800,000 dollar and another five million dollar in expected increased future revenues. By keeping track of data at the customer level, together with their purchasing patterns, Etsy can allocate its resources to improvements that have the largest impact.
Direct customer contact
Each organisation, from start-up to multi-billion corporate should understand what customers truly value. What do they want? How do they want it to be delivered? Why do they want the things they want? A problem with large companies is that only a fraction of the employees is still in direct contact with the customer. Additional reporting lines gradually blur the management’s view on the true customer needs.
Employees of successful start-ups have – due to absence of management layers – more contact with their customers and therefore know them better. Consequently, they are significantly better in delivering what the customers want. A word of advice for the large corporates is therefore to get out of the building (figurely, but also litterally). This does not mean organizing workships with internal stakeholders on what customers want, but obtaining information directly from the customers themselves. Customer behaviour and feedback should determine the company’s agenda. It is recommended to do this as early as possible in the innovation process. In times of rapid change, it can be very costly to spend a lot of time on the development of a product nobody is waiting for.
There are several ways to obtain customer insights: personal contact, product reviews, responding to social media, small surveys, showing prototypes, etc. The need for being close to the customer is not limited to one layer or department of the organisation such as customer support or sales. All employees, including those in operational functions like marketing, production or IT, should regularly be in contact with (potential) customers to understand why they are customers, or why not (yet).
Normally, teams that develop software are hardly in contact with customers. However, the ING banking app was developed by a multidisciplinary team, of which virtually everyone was in contact with the end user. Although the first, minimalistic version of the app worked fine it had only a limited number of functions. Consequently, the initial release was only moderately rated by users in the App Store. While further improving the app, the team solely focused on the customer’s reactions and ratings in the App Store. Because of ING’s customer focussed improvements of the app, the app is currently the best-rated and most downloaded app for mobile banking (Spelier, 2013). ING’s operations characterise how large corporates can derive new techniques from successful start-ups: they innovate efficiently by releasing a first version quickly to the market, they accept the risk that this version will not be perfect but they will – implicitly – promise customers to continue improving based on customer demands.
- From fully specifying before release to quickly launching a MVP and continuously improving
- From solving existing problems to testing new ideas in practice to validate hypotheses (and to pivot away from the ones that do not work)
- From working in silos to multi-disciplinary teams
- From management control to teams solving problems autonomously
Prioritising ideas based on data:
- From solving every problem to solving the most important problems
- From implementing the seemingly best solution to validating solutions by using a hypothesis-driven approach
Direct customer contact:
- From third-hand customer insights to direct customer contact at all levels of the organisation
- From risk avoidance on customer needs to willingness to make mistakes and to quickly learn from customer feedback
An article from Eric Klaassen, founding partner at BLOOM, an Amsterdam based consulting firm. He is specialised in company growth through improvement of the online business-strategy.