When it comes to really understanding how sales are being impacted by changing consumer behaviours, many retailers are turning to their data to help answer their key business questions and guide their strategic planning activities. Determining the best product range to offer your customers, or building your promotional plans for the next quarter are all vital questions to be answered, but how do you sift through terabytes of data often stored in disparate silos to get the answers you need in a timely and easy to use way? In this article, we explore some of the technology considerations of whether to build your own analytics platform or to purchase one.
When it comes to defining the requirements of an analytics solution it is easy to get caught up in the dream of having all of your data at your fingertips. There is a belief that with AI algorithms churning through all of your data to find new insights, this alone will transform the way you run your business. Too many times we have seen companies try to shoehorn all of their data into a single place without really thinking about what they actually want to do with the data once it is there.
When considering a new analytical solution, it is worth considering your business priorities. Start small, and grow from there. Are your requirements REALLY that bespoke that requires a custom-built solution? Some situations most definitely require a custom-built solution. Consider disruptor companies like AirBnB, who have had to create solutions to support their business model as there was nothing on the market that could do what they wanted. Or the autopilot software for Tesla cars. However, most business requirements are not as complex as they might first appear, and many off-the-shelf solutions will get you a long way towards what you really need. Read on to find information on the key things to consider when making a decision on whether to buy or build a new analytics solution.
Building a solution from scratch is often more expensive than people realise. Good analytical products can cost millions to build in-house, especially if the team creating it are not experts in the specific nuances of your industry. 3rd party companies who develop software for specific niches in the industry know a lot about what their clients need and can work closely with their clients to deliver great solutions out of the box. That being said, there are some great products and technologies that are designed to support a wider audience that can be used to build your own analytical solution. Companies such as Oracle, Microsoft, Google and Amazon all provide the building blocks to help you build a fully capable analytical platform.
Control of cost is also important. With a commercial solution you have greater sight of costs and can budget accordingly, even with subscription-based suppliers. With an internal team there are often more unexpected costs that could be incurred, especially in the development phase. It is easy to accidentally leave a multi-node compute cluster on over the weekend drawing down on your technology spend, believe me, we've seen this far too often!
Maintenance and new feature development are also important considerations when deciding whether to build or buy an analytics platform. Analytics technology companies are able to continually develop and enhance their product offerings to keep them competitive and pass on these new developments to their clients. These new features are often born out of industry trends and needs of other similar businesses, so when you buy a solution you get the benefits of new functionality without the need to retain a full product design and development team.
Many commercial analytics platforms already have connections to other commonly used systems straight out of the box. For example, many analytics tools can connect into various sources of information like CRM systems, POS or supply chain systems. Building these interfaces can be tricky, especially as the interfaces for many external systems change over time and can cause your systems to stop working. However, if there are a lot of non-standard or legacy components in your estate it might be worth building custom interfaces (see earlier on requirements).
When you think of commercial software products you expect them to be reliable, otherwise people will not buy them. 3rd party applications and platforms are rigorously tested before being released. Behind commercial analytics platforms there are often teams of experts keeping an eye on performance. Although this helps with the confidence that the platform is going to provide trusted insights when you need it, with it comes a greater cost. Particularly if you already have strong technical skills and trust in the quality of your source data this may be overkill for your needs.
It is worth considering that if you decide to buy a commercial analytics platform you are heavily reliant on that partner. You run the risk of products no longer being supported or functionality being removed that you have come to rely on. Companies also go bust, are bought out, increase prices, and change their business models which may no longer suit your needs. When this happens, it can take a long time to unpick their technology from your environment. When you have business-critical systems being hosted by an external company it is important to have a clear exit strategy before you enter into any agreement.
It is also important to understand the analytics that is being performed by your platform. Analytical models may not always return the same exact answers each time they are run, especially in the world of artificial intelligence. Whether you have a team of data scientists or an engagement team from an external company it is important that you understand the ‘why’ of the outputs the platform creates.
Ultimately, the decision to buy vs. build depends on your customer analytics requirements and the resources available to fulfil them. Retailers with exacting needs and a clear vision on how to reach their goals, may benefit from an exclusively built solution. While those retailers who are looking for a helping hand, maybe better served through a global product and third-party support.
If you’d like to find out how we can help with your pricing, contact us now.