Build or Buy – 8 Reasons Enterprises Leave Homegrown Integrations
Businesses with strong IT departments often look to a DIY approach for their integrations.
Writing your own integrations – or having a consultant do so – can seem like a less expensive way to connect systems. But is it?
The custom-built approach can work for smaller, low-volume integrations. But there are a number of reasons that enterprises leave them. Here are eight of those reasons, along with stories that illustrate each one.
1. Avoiding Performance Impacts on the Platform
As long as a company stays small, their minimal volumes of data delivery over web services can avoid impacting the performance of ServiceNow. But what company is trying to stay small?
Growing businesses that run reports do so on increasing amounts of ServiceNow data. And it’s not long before the company-wide impact on ServiceNow performance causes company-wide frustration.
CDW’s earliest ServiceNow integrations were based on web services. “That worked okay for a little while. But those ETL’s put a lot of stress on ServiceNow,” says Paul Liesse, Supervisor of Managed Service Applications at CDW. “I think anybody who’s tried to do bulk web services against ServiceNow knows that at some point you’re just going to run into a scalability issue.”
2. Clearing the Bottlenecks (Scalability)
When a business opts for web services for their integrations, they face several limitations. Max rows per query, batch processing, memory limits, user accounts required – any of these can create a bottleneck that keeps the company from scaling. If there are multiple connections, they may start simple, but they grow into an “integration spaghetti” that rapidly becomes difficult to manage.
ServiceNow needed to connect their ServiceNow platform to multiple SAP HANA databases. “We initially started off our journey with an ODBC connector that we provide out-of-box,” says Venugopal Malyala. “But we were not able to scale and provide real-time integration.” They found a solution that gave them the necessary throughput – now 13 million+ transactions per day from 600+ database tables. “Our biggest challenge was the volume of the data that we are processing.” With a new solution, “we were quickly able to scale.”
3. Reducing Maintenance
DIY integrations often require more maintenance than a company initially plans for. New privacy rules come into effect. The integration breaks during an upgrade. A developer leaves – and takes their knowledge with them. It’s difficult or impossible to monitor transactions or to recover lost data. These and other problems often require more attention than a company has the means – or the ability – to handle.
Zurich Insurance faced the need to integrate their global helpdesk (running ServiceNow) with multiple service providers. They initially used web services and APIs, but they had difficulty supporting those integrations. When a transfer failed, they didn’t have a way to monitor why it failed or to recover what data was lost. Realizing that the homegrown integrations were too difficult to maintain, Zurich Insurance started shopping around.
4. Upgrading Systems Without Breaking Connections
When a business needs to update ServiceNow or another system, they get nervous about the custom development on that system. When the system updates, homegrown integrations usually break – and need to be rebuilt. Dozens or hundreds of hours in development labor is lost.
When Virteva (now Crossfuze) signed on for a data-automation solution, they started integrating more on the process level. When they upgrade a system, it’s just a new endpoint. The common data model in their solution makes use of pre-built mapping. Because Virteva integrates at the process level, they can change systems or upgrade with confidence, because the automation solution has no problem with endpoint changes. According to Matt Miller, VP of Delivery at Virteva, “I don’t have to carry that burden to make sure that I’m compliant with every version of ServiceNow – and make sure that I’m complying with every version of JIRA and Salesforce.”
5. Making Advanced Reporting Easy
To take advantage of the powerful features of advanced BI tools, enterprises extend their ServiceNow data to a database. But reporting on that data will be useful only if the data is there. And it’s hard to get all the data there with web services – without crippling the platform in the process.
Intermountain Healthcare needed the kind of advanced reporting on ServiceNow data that was available only through a powerful BI tool. “I think there comes a point when we decide, how much time do we want to spend writing this stuff versus buying in an external tool…. This may be possible other ways. But I tell you what, with Perspectium and a local database, it became very easy.”
6. Avoiding Data Loss
When a company builds an integration with web services, they rarely have any way to verify that a transaction takes place. And data loss – without possibility for recovery – is an ever-present risk.
At one time, Virteva (Crossfuze) was using web services for their integrations. When they did so, they had to manage the headache of lost data and hours of troubleshooting. “We weren’t getting all the updates from our customers’ instances,” says Matt Miller. Virteva was “working endless hours on troubleshooting where something left one instance but never showed up at the other, and understanding where that gap was, and what we could do to solve that problem.” Eventually, Virteva signed on with a vendor that provided data queuing and ensured that transferring data was still available – even in the face of an endpoint outage.
7. Relying on Expertise
After the business world started opting for software as a service (SaaS), businesses eventually realized that they could take the “as a service” benefits and apply them to other areas – including integration. Integration as a service shifts the burden from the in-house development team to the vendor. This provider has expertise in security and privacy rules, in the platforms to be integrated, and in the ability to ensure the performance of the platform and the connection itself.
Accenture faced the choice of whether to build or buy. Among themselves, they hashed out what the ideal solution would look like – and they found it, without needing to build it. “If we were going to build it, this is exactly what we would have built,” says Jonathan Livingston. “And to have someone already have built it and have the expertise already in the tool out there, that was that was a big find for us.”
8. Saving Time and Money
Eventually, the numbers just add up. Homegrown integration projects frequently go over budget. Unexpected maintenance. Platform changes that break the integration. Chasing compliance for security and privacy rules. Accruing developer hours. All add up, and the total cost of ownership at a growing company becomes a convincing case for leaving homegrown integrations.
Fujitsu, a managed service provider, decided that a new integration with a customer took too long and cost too much. And this problem repeated every time they wished to make another connection. Eventually, they signed on for integration as a service.
The Modern Approach to Data Automation
For companies outgrowing their standalone integrations, an advanced approach to data automation just makes sense.
Such companies work with a vendor that can implement quickly, using standard deployment with in-built best practices. At Perspectium, we take responsibility for the data automation, ensuring that the connection continues working securely and reliably.
Want to learn more about the modern approach that replaces homegrown integrations? Let’s chat.
Published Date: September 10, 2020