Fair warning: This is more about not having written anything in a while than the value of the topic…and the subject matter is more about drawing your own conclusions than relying on what is easily available, so…
App is one of the most over-used and ill-defined terms in the IT lexicon. This is greatly due to it being used by people outside the IT domain. The domain itself has had some whoppers, like the DHMTL that was a must-have at the turn of the century even though the only honest definitions of the term were that it had no real definition. Microservices runs a close second simply because there is an invisible grey line between SOA and Microservices that is a mile wide and an inch short. But I digress, as is often the case.
What I’m really thinking about today is apps in the world of Salesforce.com. Specifically, apps that run inside the Salesforce CRM platform. I started thinking about this because I was looking into CPQ vendors over the weekend to refresh myself on the market to support a project proposal to select the best option for a particular business. It’s a large space, so it always helps to find someone else’s list to start with and someone had given me a list from a major analyst group as that starting point.
Other than analysts, no one likes long lists with lots of details, so I first wanted to narrow it by those that integrated with Salesforce. It didn’t take me long to remember that Salesforce is the gold standard for CRM and there were only two that didn’t. I didn’t go through the whole list to get to that count because I’ve done these kind of evaluations before and figured out after the first half dozen that this was not how I was going to narrow the list. The two were just what was noticed while skinning this cat another way.
The first trimming of the list was by industry focus. The potential client is a tech service, sort of SaaSy, and “High-tech products” was one of the categories, which was much closer to what they did than “Financial services” (though they have customers in that domain) or “Industrial products” (which the analyst seemed to think usually included high-tech, though not sure why).
To spare you the tedium of the several hours of wading through 1000’s of lines of marketing prose that could have been delivered in a table (ahem, yes, I know, kettle, black, etc.), from just the perspective of Salesforce CRM integration, I found it useful to divide them into three basic styles:
Native: An application that is built entirely in Salesforce
App: An app that runs inside Salesforce that depends on data and/or functionality managed outside of Salesforce.
Connector: An application that runs independently of Salesforce and has a way to share data with Salesforce.
The terms for these distinctions change often over time and between sources. These definitions are for clarification of the table below and are purposely simplified as deeper distinctions are less relevant about integration than other aspects.
In this particular exercise, the ask was to provide some pros and cons to these different styles. My style being one of adapting general terms to technical solutions, I responded with a non-exhaustive list of Benefits and Concerns:
- Easily accessible in the sales process context.
- Seamless integration with other native apps.
- Has gone through Salesforce security review.
- No data latency.
- Easily accessible in the sales process context.
- Access is managed within Salesforce.
- Has gone through Salesforce security review (only if installed through App Exchange).
- Control over storage impacts.
- Broader range of vendors to choose from.
- May require additional Salesforce licensing.
- May have impacts on storage limitations.
- Frequently limited functionality.
- Support may require coordinating the vendor and Salesforce.
- High potential for latency.
- Difficult to trouble-shoot.
- Users must use multiple applications.
Of course, the next question is usually “which is best”, and I must respond with the architect/consultant/writer-needing-higher-word-count with “it depends”. And it depends on lots of things, such as who will be maintaining the solution; how are capex and opex prioritized and managed; how do different stake holders actually need to interact with the solution; and is it clearly understood that this only one aspect of a vendor selection process and all known aspects must be documented and weighted before giving a recommendation?
The real reminder for me when I finished this brief analysis was that context is everything when doing any type evaluation. The list that I started with included products that were questionable as to whether they really belonged in the report and many of the products were listed as serving domains that there was no mention of on the vendor’s site and no compelling reason why the unmentioned-domain would want to use it. If I had direct access to the author(s) I may learn something by asking, but the important thing is that I used their input only as a starting point and applied my own analysis because when the recommendations are provided to a client, those author’s name will not be on the agenda and they will not be there to answer the questions that hadn’t yet been thought of.
© Scott S. Nelson