Reality check on your Salesforce data integration project

If you're embarking on a Salesforce data integration project, you shouldn't forget these three overlooked considerations, expert Eric Borthwick cautions.

Most enterprises choose software based on a tool's ability to solve a business need. This choice often has to occur...

within a company's existing parameters: Does it have the budget, in-house skills to maintain the software and the current technology portfolio to ensure that new technologies can integrate with existing systems?

Salesforce, the cloud-based CRM vendor, is attempting to make the need for data integration a thing of the past. It has designed its broad-based platform to integrate with numerous applications, both native and third party, based on APIs or connectors that join these apps and their data together. Salesforce does this because it recognizes that a significant portion of its customer base won't be able to move legacy systems in the short term or may have installed competing products. As a result, in 2014, the CRM provider released Salesforce1 Lightning Connect.

But the notion of turnkey data integration is still fictional. There are concerns in any integration project. And in the case of Salesforce data integration, admins needs to identify the integration points between systems; determine how often data sources should be in sync; consider whether the data needs to be moved bidirectionally; and, finally, map out whether it has the tools and staff necessary to complete the project. Answering all these questions will help position an enterprise to deal with its single largest concern: cost.

Considerations in integrating data

If your company is trying to evaluate the cost of Salesforce data integration -- from the staff time to your ability to integrate legacy systems and the nature of the data you want to integrate -- the following factors should be considered:

Unfortunately, many integration projects are improperly implemented.

Real-time vs. batch integration. This is a good place to start thinking about these kinds of questions. It will help determine the cost of licensing an integration tool. Making the determination is fairly straightforward. For instance, a company selling tens of thousands of tickets to a stadium, which might sell out in a matter of minutes, needs up-to-the minute data accuracy among systems. To ensure that it sells an individual ticket only once, real-time integration is required. Conversely, a yacht manufacturer that sells only a few units a year might want to integrate data on a daily basis; its planning needs don't need to be up-to-the-minute, and batch integration is a better fit. Generally, batch integration tools are significantly cheaper than real-time technologies. If considering a batch integration solution, the user should be aware of scheduling, the amount of time covered within the batch and how logic within the processes is configured.

Number of integration points. The second concern is the number of integration points. While many tools charge by the integration point or have a tiered pricing model based on thresholds of integration points, one must also consider the length of time to configure each of the disparate systems and time to build logic for the integration tool. In most systems, naming conventions don't line up, or necessary integration fields may not exist. Creating mapping to show the destination of the data is a necessary function of any integration project. And maintaining the systems, naming conventions and configurations after the integration also must be considered and well documented.

Bidirectional integration. Bidirectional integration allows for two integrated systems to have data entered via each independent system, but still be in sync. This happens often in ERP-to-Salesforce integration. Many users will change pricing or pricing models within the ERP and want this data to update the price books within Salesforce. At the same time, the sales orders completed within Salesforce would need to be synced with an ERP system to have all up-to-date orders for invoicing purposes. Cost for bidirectional integration can be calculated with a simple count of integration points in both directions.

Salesforce1 Connect was released with a licensing price of $4,000 per month depending on data source. This pricing reflects only operability for real-time Salesforce data integration. The pricing seems roughly in line with other products in the marketplace. Batch integration solutions are available from many other providers. The time to configure a technology also depends on the experience of the team members in integration and the depth of their understanding in integrating disparate systems. Testing environments for all systems involved are necessary to protect business continuity.

Unfortunately, many integration projects are improperly implemented. Most companies aren't prepared for the kinds of questions above and don't ask them. Without this kind of upfront planning, though, data integration projects are doomed to a false start. With the correct planning, the benefits can certainly outweigh the risks.

Next Steps

Mobile CRM is all about data integration

Learn how BPM helps data quality

Why clean data is essential to sales strategy

Dig Deeper on Customer data management