Moving an e-commerce platform is generally a very big deal. Most retail or B2B organisations already have a fairly mature e-commerce business, and therefore a re-platform has the potential to cause a big disruption to that business. The e-commerce operation is likely a significant driver of to both revenues and profits to the bottom line, and getting it wrong can have a major financial impact.
Over my career, I’ve been involved in many e-commerce re-platform projects, some small and some very big. There are projects that have gone pretty smoothly, while others have been a bit bumpy – but there’s often a common thread through many of them: There’ve been a number of areas that have been underestimated or under-appreciated by one degree or another which have had an impact on the success of the project.
Planning an e-commerce re-platform: 9 keys to success
Let’s review the factors that come into play when planning an e-commerce re-platform:
- Business readiness
- Discovery and scope
- Agile development
- Data migration
Ready? Let’s get started on the details.
We all know how important SEO is to e-commerce business. Companies can live or die by their organic search traffic, and a big dip in this can be very expensive, especially if the gap is filled with paid search traffic.
While everyone knows SEO is super-important, it’s often one of the last things that is considered during a re-platform. Businesses tend focus on their functional requirements, UX, data, and integrations, and will often only involve their SEO agency towards the end of the project once the new platform is in the testing phase. This can be a big mistake that can delay a project and significantly increase costs.
A re-platform will often involve a wholesale change to content, product data, UX, and URL structure – all of which can have a big impact on SEO.
While it’s always expected that a website will see a dip of up to 20% in its inbound organic traffic after a re-platform, involving your SEO agency at the very beginning of the process can help limit that impact and recover any lost SEO equity as quickly as possible. I’ve certainly seen instances where an SEO agency has had sight of a new platform late in the development process and the business has had to make the choice of going live with a platform that’s not fully optimised or delaying the go-live in order to implement the SEO recommendations before going live.
I’ve previously written articles on testing and how important it is, but I would still class this as one of the biggest gotchas in a re-platform project. The testing effort required to get a re-platform project complete and ready for go-live can be immense.
A business will have spent many years ironing out all of the kinks in its legacy platform and may not appreciate the fact that it has a completely new platform that needs a huge amount of testing. The platform may have:
- Multiple payment journeys
- Many integrations
- Complex promotions
- A lot of data migration
- An entire new tool-set for the business to use
There’s always a tendency to try to squeeze the UAT phase down, as everyone wants the new platform to go live as soon as possible. It may have taken 9 or 12 months (or even longer) to build, which is a long time to wait, but skimping on testing is only going to result in a lot of pain post go-live.
The reality of software development is that a lot of defects will occur and these will only be found through thorough testing. In the past, I’ve seen a lot of ‘happy path’ testing where testers will simply follow a simple user journey testing only the basic scenarios. This may be due to inexperience or due to time pressures.
Proper testing involves trying your hardest to break the system and can take time. Multiple edge-case scenarios should be tested as the sum of all edge cases add up to quite a lot. You can guarantee that your users will do everything imaginable with your website, double-clicking, hitting the back button, entering in the wrong data, and they will find every defect there is. Going live with an e-commerce re-platform that has not been sufficiently tested is always going to hit your customer’s experience.
We will always recommend that at least 2 months is planned at the end of development for UAT – even 3 months if you want to be thorough. This may sound like quite a lot, but if testers have many hundreds of test cases to run and any defects found need to be fixed and retested, this time can be used up very quickly. Many companies will budget for a testing effort that is anywhere between 20-40% of the total development time.
You should not just wait until the end to test everything. Building in a level of automated testing from day one, as well as system testing between each development sprint can help to de-risk the project and reduce the number of defects that are found during the final UAT phase. This does take investment up front but, if you don’t test early, you will end up having to all of the testing and fixing later on which can be much more painful.
Business readiness: If you think you’re ready for the e-commerce re-platform, always double, triple, quadruple confirm
Business readiness is, by far, the biggest gotcha of them all. An e-commerce re-platform is a very big undertaking for any organisation and involves a huge amount of effort and change on the side of the business, which all needs to be managed very carefully.
The business is likely to already have an e-commerce operation, so it’s unrealistic for those staff members who are busy trading that e-commerce platform on a day to day basis to also manage the re-platform. It’s very easy to underestimate the amount of time the business will need to dedicate to the re-platform throughout the entirety of the project.
It’s important that the business works with its development to develop a strong business readiness plan. This should be done at the start of the project so that the business knows exactly what time, effort, and skills it needs to dedicate to the project during development, during go-live, and then afterwards to ensure that the project is a success.
Discovery and scoping
The discovery phase of a re-platform can be a very intensive process and requires the involvement of many different stakeholders within the business. It’s not uncommon for us to run up to 30 different discovery sessions each involving up to 20 different people from the business.
This all adds up to a lot of time and effort and can take these people away from their day to day jobs. However, this is an essential part of the project, as it’s effectively creating the blueprints for what is going to be built. Without it, the business will not get what it wants, and the project will very rarely be a success.
Most re-platform projects are run using a hybrid waterfall / agile methodology. The project normally starts off as waterfall during the discovery / scoping phase.
Once the requirements have been gathered and understood, a design and plan is formed. The requirements are grouped into epics (high-level topics) and the development and implementation plan is split into small sprints, each containing specific epics.
The benefit of running the development phase in an agile way is that it gives the business visibility of what is being built very early on and allows it to test whilst the platform is still being built. This model gives the business the flexibility to change the scope during development but also helps us ensure that the product delivered is just right for the business.
However, involvement in an agile development programme can require quite a lot of time from the business. Sprints are usually 2 weeks long and have various ceremonies that happen within each sprint, such as daily stand ups (meetings) and user story elaboration sessions which can take all day.
The business will need to have a number of staff who are dedicated to the project during this time otherwise the velocity and quality of the development will be impacted.
Data migration: You can’t have a successful e-commerce re-platform without it
This is a biggie. Depending on your business, you’re going to need to migrate the following from one platform to another:
- account data
This can be a major undertaking and we normally advise our clients to take a pragmatic view on data migration.
Do you really need to migrate the account of a customer who last bought from you 5 years ago? Regardless of how pragmatic the business is, migrating data from a legacy platform to a new one requires a lot of effort on its part.
There is undoubtedly going to be a data cleansing exercise that’s going to happen, and the business is also likely to use this as an opportunity to tidy up its product data and classification. In my experience, this is almost always underestimated and should be started at the very beginning of the project with a lot of time dedicated to it.
A business is likely to have a lot of content within its legacy platform, but this often needs to be re-created for the new platform. The UX and page structure is likely to be different and therefore much of the existing content cannot be easily slotted in.
It’s easy to underestimate the amount of time it takes to create the content that is required for a new platform. You always have more content than you think.
It’s strongly advised to put together a content creation plan at the very beginning of the project so that you have full visibility of exactly what content needs to be created to ensure that it’s all ready when needed.
Training: Your e-commerce re-platform is live – does everyone know how to use it?
A re-platform involves a business learning to use a whole new set of tools from:
- order management
- promotions management
- content management
- customer service management
A mature e-commerce business is likely to have many different people that will use the platform. Ensuring that everyone is sufficiently trained in how to best use the platform is a big undertaking, but is vital to get right.
It’s important to train business users at the right time. If you do it too early, they’ll just forget what they have learnt by the time the platform goes live.
While a large part of the re-platform is a technical change, it also involves a lot of business change. Ensuring that business users are sufficiently trained in the new processes as well as how to use the new technology is vitally important.
You can pretty much guarantee that some things are going to go wrong during the go live phase. They always do. It’s really important that the business has the right reporting tools in place to give everyone visibility of exactly what’s going on.
You may have multiple, different ways your customers can buy, and if you see a dip in your total sales or conversion rate after go-live, you need to be able to quickly understand where that dip is coming from. One of the best things about e-commerce is that pretty much everything can be made visible and be analysed.
There’s really no excuse for not having all the data you need available to you as it is all there. During the hectic post go-live phase, everyone will be stretched and there will be a lot of noise, so being able to quickly determine where issues lie and then being able to focus your efforts on dealing with those issues will be vital.
Overall, an e-commerce re-platform can be a daunting prospect and can have a huge positive or negative impact on a business.
Ensuring that a business has full visibility of the investment that it will need to make in people, time, and effort is important for its success. The business should start to put together a business readiness plan as soon as the project starts (or even before then if it is very keen), as the business’s role in the project is just as important and as involved as that of the development partner.