6 Hidden Costs of DIY ServiceNow Integrations

When a company builds their own integration, it is natural and obvious to consider the building cost. But there are other less obvious costs to DIY ServiceNow integrations. In general, companies incur these costs after they build the integration.

A previous post discussed how rising turnover of personnel increases the cost of custom-built integrations. Let’s consider 6 additional hidden costs to DIY ServiceNow integrations.

1. Testing and Quality Assurance

After building the integration, it is important to test it under a variety of conditions. The hidden costs here are primarily in the hours of work.

• Is data arriving at the other endpoint?
• Are certain errors common? How can you reduce them?
• Can the integration handle large volumes of data?
• If necessary, does the integration adequately flow data bi-directionally?
• Is related data transferring?
• Does the integration capture and transfer updates to data?

These are just some questions to consider during testing. As you consider hidden costs throughout this article, incorporate their consideration into quality assurance before deploying the integration. From survey data on building ServiceNow integrations, we estimate that on average, 100 hours for quality assurance and testing are necessary. 

2. Integration Failures

As much effort as you expend on quality assurance, it’s important to anticipate the cost of integration failures. This consideration is often missing. “No one thinks necessarily about the technology to deal with a failure,” says Carey Blunt, who manages multiple integrations at Fujitsu.

An integration failure refers to a non-existent or incomplete transfer of data. These costs are more difficult to calculate, but you can still estimate them.

Cost of Missing or Inaccurate Data

Sometimes an endpoint experiences an outage. Other times, the endpoint is so busy with other activity, that the transfer timed out and failed. 

There is a reason that you set up an integration between the two endpoints. It must offer some value for your company to synchronize that data. How much value is lost when the sync fails?

• When the data does not flow to the intended endpoint, what is the cost to your company of not having those updates there?
• If the data did not arrive at the data warehouse, what is the cost of missing or inaccurate data there?
• If the data did not arrive at the other ITSM app, what is the cost of that obsolete ITSM incident data?

Flawed data also risks the timely delivery of services and jeopardizes SLAs. In general, bad data across the organization costs companies about 20% their Revenue. Integration failures certainly contribute to that cost.

Cost of Recovering Lost Data

When the integration fails, sometimes the failure stays undiscovered for hours, driving up the cost of missing or inaccurate data. It may be the integration team that discovers the failure. More often, it is those who work with the data who discover that it is missing or inaccurate. They then alert IT.

“Endless hours on troubleshooting.” This was the maintenance experience at Virteva (now Crossfuze) while they were recovering lost data when web services failed for their ServiceNow integrations. Matt Miller, VP of Delivery, tells the story:

“We had the REST integration between ServiceNow and ServiceNow. It was functional. It was not the best solution that we thought was possible at the time.

“We were having challenges with data drops between rest calls between instances. So we weren’t getting all the updates from our customers’ instances.

“Working endless hours on troubleshooting where something left one instance but never showed up at the other and understanding where that gap wasn’t and what we could do to solve that problem.“

To discover the cost of recovering lost data, add up the hours for this kind of troubleshooting.

Cost of Monitoring for Failures

After repeated failures, usually at a higher-than-expected rate, IT decides to monitor the integration more closely. Such monitoring reduces the cost of integration failures – but is itself a cost.

3. Continuity Through ServiceNow Upgrades

During a ServiceNow upgrade, any integrations via web services must pause in order to avoid data loss. This cessation could mean running the upgrade during off hours – and hoping it will complete before business resumes.

Alternatively, you can log all updates made during the upgrade and sync them after the fact. Either way, without data queues in place, there will be time and cost devoted to monitoring and maintaining integrations.

After the upgrade completes, test the integrations before business resumes. More than likely, an upgrade will break integrations, which will need to be rebuilt.

So, possible integration costs around upgrades include …

• Preparing for the upgrade by pausing integrations
• Logging all updates during the upgrade for loading once the upgrade is complete
• Rebuilding and testing integrations that the upgrade broke
• Working with missing or inaccurate data until updates are loaded and integrations can be rebuilt

In the face of a ServiceNow upgrade, a custom-built integration can start looking a whole lot like technical debt.

4. Ongoing Training in Best Practices and New Technology

Vendors who offer packaged integration processes stay up-to-speed on latest technologies and best practices involving integration. In addition, they are in themselves a network of experts that help one another flourish in their work. 

At companies using DIY connections, an integration admin takes on this training themselves and needs access to a network of skilled integrators that can offer help on individualized challenges. Keep in mind turnover rates and the possible need to repeat this training.

5. Performance Impacts Due to Integrations

For integrations that transfer limited amounts of data, a custom-built ServiceNow integration often makes sense initially. As the data volumes grow, however, the impacts on ServiceNow performance add up.

When an organization uses web services to connect ServiceNow to other apps, they add traffic to the same channels that their ServiceNow users operate within. During these queries, the impacts to ServiceNow performance reduces efficiency across the organization.

In other words, ServiceNow users company-wide must wait longer for their ServiceNow actions to complete. And as they wait, they often refresh, adding more impact to ServiceNow performance.

These performance impacts are not exactly hidden – they are painfully visible. But the costs are, because the performance impacts can be challenging to quantify.

6. Delayed Time to Value, Due to Delayed Implementation

Building, testing, and implementing a ServiceNow integration usually takes longer than expected. Though the average time to do so is 470 hours, many integration builders anticipate a faster implementation – and then incur costs due to delays.

Before signing on with Perspectium, one insurance company spent 6 months and $400,000 on attempting to integrate with a service provider. Incidentally, they never fully deployed these custom-built integrations, which is why they came to Perspectium.

Whatever the business imperative for the integration, a delay in implementation means that the business must delay the value it plans to realize from the integration.

Costs that arise from this delay can include …

• Continued inefficiency and risk of error in swivel-chair work
• Lack of business intelligence
• Flawed decisions based on reports reflecting inaccurate or obsolete information
• Delay of advanced initiatives (e.g., AIOps and other machine-learning projects)
• Direct revenue hit for MSPs that delay onboarding of customers

When calculating your integration costs, consider adding the possible costs of delayed implementation.

Get a Bigger Picture of Total Cost of Ownership

Along with staff or consultant turnover, these hidden costs of custom-built ServiceNow integrations contribute to the total cost of ownership (TCO).

Hidden costs undoubtedly make the calculation of TCO more difficult. To simplify the process, try out our integration TCO toolkit.