In today’s data-driven world, avoiding data silos is vital for efficient workflows involving data-generating departments – which is exactly where data integration comes in. But how do you decide on an integration strategy? Is it possible to just pick a tool and go with it? Many businesses do this, but there might just be a better approach.
Why is data integration vital for your business?
When your employees use software systems – such as your CRM and ERP systems – independently from each other, they’re probably constantly switching between the two in order to collect all the information needed to do their jobs.
When drawing up your quote, if you have to look up customer details in the CRM system, then search for product information and an up-to-date price in the ERP system before you can even start, a lot of time – and money – flies right out the window.
Sharing data between your enterprise applications gives your users total visibility over relevant data and provides actionable insights across these applications. This makes your business processes more efficient, reduces the risk of errors, and enables better and faster client service, which has a direct effect on your bottom-line.
This doesn’t mean that you always need a full-scale integration to achieve your goals. Sharing data can also be achieved by for example a visualization of that data. This alternative to an integration might also pay off.
How to get started?
To choose the best data integration – or visualization – strategy for your business while controlling costs and ensuring stability, agility and security, take a step back first. Examine your processes, services, applications and data to clarify the design and requirements of your ideal integration.
Many software developers currently standardize integrations as much as possible. Although this can be interesting, it often requires your applications to use a standard setup as well.
Custom integrations, however, don’t necessarily result in high costs, thanks to the rise of cloud infrastructure and services. If you take a proactive approach to your business processes and use your business data to automatically schedule ‘next best actions’ (for example, triggering an e-mail, renewal or reorder), you may eliminate the need for complex data integrations alltogether.
A standard integration tool won’t be a right fit if your business processes, ERP and CRM systems aren’t standard either. Costs will escalate quickly if you have to adapt your solution to suit your needs.
Step 1: Map out data flows
First, map out how and when data flows across your applications and business processes. Identify the data you need from each system and why you need that data for operational processes.
Step 2: Define integration requirements
Define the degree of data exchange necessary between the systems. Do all of your systems need full access to all data in real time? Is it necessary to be able to manage and edit this data, and enable updates across platforms? In many cases, data visualization – via a dashboard to consult ERP or CRM data, for example – is already enough.
When your marketing team sets up an automated e-mail flow triggered when new customers are entered into the ERP system, they probably don’t need to adjust customers’ address or financial details in the CRM system. You might actually hope that they don’t, as it could result in conflicting billing details in the ERP system.
Getting integration requirements clear from the start will allow for a less complex integration strategy. Why integrate everything when doing so isn’t necessary to fully support your business? Limiting integration to your urgent needs only enables you to dodge difficulties that could arise when merging data from applications with different architectures.
2 data pitfalls to avoid
- Different data models: when two systems process and store data based on different data models, merging them can be a challenge. An example: if client details are represented in the ERP system in 10 fields, extensive mapping will be needed to align this data with a CRM system that structures that data in 5 fields.
- Unstructured data: when databases are fed manually with information, inconsistencies such as spelling mistakes, formatting variations and differences in naming conventions can delay integration or even make it impossible.
However, many enterprises skip steps 1 and 2, immediately pick a data integration tool – preferably the most standard one possible – and begin their full-scale integrations. Unfortunately, the choice of a standard tool often leads to the problems above, leading many business owners to believe that data integration is unstable and unreliable. But in many cases, they just didn’t choose the right tool for their project.
A standard integration tool won’t be suitable if your business processes, ERP and CRM systems aren’t standard either. Costs will escalate quickly if you try to adapt a standard tool to fit complex requirements.
Step 3: Choose your integration tool wisely
However, if you’re going for a standard integration between two standard software systems, it’s ok to rely on standard ETL tools.
ETL: extract, transform, load
An ETL tool integrates data from multiple applications in three phases:
- Extract – it extracts data from the source system.
- Transform – it transforms the extracted data into a format suitable for the target system by applying a set of rules or functions which (should) solve the merging problems mentioned above.
- Load – it loads the new data into the target system.
As soon as your integration deviates from this standard scenario, you’ll quickly experience the limitations of these tools. It can be difficult to pinpoint the exact reasons why problems occur, especially if you lack technical computing knowledge.
Moreover, these integration tools often form tight connections between two systems through an end-to-end architecture. This means that if something goes wrong in platform A, because platform B is entirely dependent on it, issues cascade throughout the entire system.
Enterprise service bus
You can work your way around this challenge by choosing an enterprise service bus to integrate your systems. Instead of integrating each application directly with another, an ESB functions as a middleman that distributes needed data to relevant applications upon request. This integration method is completely automated; error logs and problem-solving functionalities are so reliable users don’t need any technical skills to work with it.
As flexible as this software architecture is, systems that use it, such as Microsoft BizTalk, can be expensive. Licenses are pricy, you need adequate infrastructure, and it is often necessary to hire specialists to set it up. For smaller businesses, the benefits probably don’t outweigh the costs.
Microsoft Azure Integration Services
Azure Integration Services is a cloud-based integration solution from Microsoft which combines a set of functionalities, including a service bus, into a complete integration solution for all systems running on premise and in the cloud.
If you’re looking for a more affordable integration method which is easy to customize to your needs, Azure might be right for you. Unlike more traditional integration software, you can choose and pay for only the Azure cloud services you need.
Since all services are readily available at the click of a mouse, integration mistakes are virtually impossible.
Conclusion: decide on your requirements before choosing a tool
Which integration tool works best for your business depends on the data you want to exchange between your systems, whether you need to consult that data or use it to manage processes and when the data has to be made available – in real time or not.
As soon as you have a clear picture of what you do with your data, it’s time to look into the integration tools available and make an informed decision. Take the time to do your homework. Don’t assume that a standard tool will be the most cost-effective solution, because in many cases, adjustments lead to costs piling up.
Looking for the best integration strategy?
Cegeka can help you at every step of the way. Let’s talk.
Rob Hufschmitt, Solution Manager