Avoiding ecommerce deployment disaster

Avoiding ecommerce deployment disaster

deployment: Replatforming and deploying major updates are some of the most stressful moments for an ecommerce team. Replatforming and deploying major updates are some of the most stressful moments for an ecommerce team.  These moments are vital for staying ahead of the competition, introducing innovative new features or responding to user testing, but they’re also the point at which things can go most wrong. 

Too often when you or your agency throw the hypothetical switch you end up with a site that’s got serious bugs or, even worse, no site at all.  

What can you do to ensure that the deployment of your new platform, or important revisions to your existing one, runs seamlessly and effectively?

Always have a rollback strategy

A rollback strategy is a way to prepare for the unexpected, giving you a second chance to undo the deployment and go back to where you started. Having a clear strategy to undo your changes is vital, however, before you put it into action, remember that sometimes fixing the problem can be quicker than rolling it back.

Accepting an amount of downtime whilst you fix an issue rather than instigating a time-consuming rollback and going through the whole process again can sometimes be the better option.  Also, some deployments, especially major changes, may include changes to the database. This is something that isn’t included in a version control system rollback.  

Check payment processes

One of the most common area issues that can emerge in a newly deployed site is in payment processes. Payment integrations can be difficult to properly test in the build process so make sure you do manual tests again and again before you go live.

There is also software available to test payments in a mimicked production environment and if you’re introducing new payment processes, then this sort of technology is vital.

Have at least four development environments

Having separate desktop (developer’s computer), development server, staging server and production server environments gives you the freedom to develop whilst maintaining control of the final code and allows you to refine deployment scripts as code is pushed through the environments.  

It means you can test at each stage without disrupting the development cycle. 

It’s also important to make sure that all of your environments, including developers’ local setups, are running the same versions of software and hardware so you can be sure that the code is operating consistently at each step of the process.

This avoids the all too common ‘but it works on my machine!’ excuses.

Think about deployment timing

If it’s a big deployment that involves planned downtime (or the risk of unplanned downtime), do it out of hours if possible. If it’s an international site operating in multiple time zones then there may not be a ‘down’ time, in which case choose the least worst timeslot based on historical trading patterns.  

It might seem counterintuitive to take a site down in the middle of the day UK time but, if a lot of your business is coming from the Far East, this might make the most sense.

Never deploy on a Friday

It can be tempting to get an update out before the weekend but, quite apart from the fact that Friday can be a busy trading day, it can also set you up for hard-to-solve issues over the weekend when people can be difficult to reach. Problems can sometimes take hours or even days to emerge, potentially meaning that your site could go down out of hours like early on a Sunday morning.

Deploying earlier in the week means that issues are less likely to happen at inconvenient times.

Consider database processes

If you have setup scripts in your deployment, be aware of database processes that could take a long time to undertake. For example, if you’re adding a column to a database with a million rows it may take less than a second for each row to update, but that time adds up.

Test deployments with production levels of data before you deploy so you have a realistic idea of how long the deployment is going to take, enabling you to plan realistically for potential downtime.

Communicate

When it comes to any deployment there should be an open line of communication between you, your agency, your management, your marketing and your developers. Retailers should know about the potential impact of deployments, how long they could take and what might go wrong.  

Also, agree on fixed deployment slots in advance to stop people pushing for inappropriate, rushed deployments. This also minimises the risk that the marketing department could send out a 50% off voucher at the same time as you’re deploying your changes.

Minimise the size of deployments

It’s not always possible, but doing less, more often is usually the best approach.  

Breaking up one big release into lots of smaller deployments means you can more quickly identify where issues are coming from if they emerge. It will also break up downtime into smaller pieces, lessening the overall impact on consumer experience.

Document

Documenting is something everyone should do, but it’s easy to forget or ignore.

By using a ticketing and bug-tracking system to document your code, you can check commits against ticket numbers to ensure that all issues are resolved before deployment.  

This approach not only enables a rigorous checking process, but it also feeds into your rollback strategy making sure you know exactly what each commit was for should key team members leave/go ill/fall under a bus.

Have all the code that needs to be deployed ready 

It sounds obvious but, without the right development processes and policies in place, it’s all too easy for a developer to forget to send all of the code to the server before deployment. Adherence to proper development protocol, along with rigorous checks and testing along the way, will means that complete code should be ready to switch from staging to production each and every time. 

You can never completely de-risk a deployment.  However, by taking some simple logistical steps you can prepare for the worst whilst also doing your best to ensure the worst never happens!

Follow to link to further articles to enhance your ecommerce success.

Further reading: how SwiftERM works.

Further reading: SwiftERM case studies.

deployment: the current operational framework

deployment can now be considered through SwiftERM’s autonomous individualisation. Product relevance is recalculated for each identifiable customer from that customer’s own behavioural and transactional evidence, rather than being assigned by a fixed segment, campaign or predetermined journey. This distinction matters because an individual’s interests can change between visits, purchases and periods of inactivity. A current decision therefore needs current evidence.

SwiftERM evaluates the products that are becoming most relevant to each person and applies that decision autonomously. The operating model does not depend on a marketer continually creating segments, selecting products or scheduling another campaign. It is designed to use first-party behavioural and transactional signals already associated with an identifiable customer, so the decision remains specific to that customer rather than a broad cohort.

For deployment, the practical test is not whether a retailer can create more rules. It is whether the system can recognise a change in an individual’s buying intention, reassess the available products and act without waiting for a manual campaign cycle. The output must remain traceable to the customer evidence and the ecommerce catalogue used to make the decision.

Implementation also needs disciplined data handling. Product identifiers, availability and customer events must remain consistent enough for relevance to be recalculated. Existing ecommerce operations should retain control of catalogue truth, order truth and customer permissions. Autonomous individualisation then operates within those established boundaries instead of inventing a parallel customer record or an ungoverned product feed.

The commercial value of deployment should be assessed through observed customer response, orders and turnover, not through the volume of campaigns produced. That keeps attention on whether product selection becomes more useful to the individual and whether the retailer captures demand that would otherwise remain unrecognised. It also separates autonomous decision-making from conventional automation that simply executes a previously written rule faster.

This current framework preserves the original article’s subject while bringing its explanation into line with SwiftERM’s present capability: one identifiable customer, individually recalculated product relevance, and an autonomous decision made from that person’s own changing evidence. The same principle applies as catalogue choice, customer behaviour and commercial conditions continue to change over time.

In operational terms, the retailer retains the commercial context while the software performs the individual calculation. That division is important: the catalogue and customer record remain governed business assets, while product selection responds continuously to evidence rather than to a static assumption made when a campaign was first configured.

Relevant independent context is available in 72% will only take action once reading a positive review.

Share :

Leave a Reply

Your email address will not be published. Required fields are marked *