Dynamics CRM Ecommerce Integration: A Better Discovery Process Before You Build
A Better Discovery Process Before You Build
Most projects start with solutions. There is a demo, competitor presentation, or vendor meeting where everything looks easy and seamless. Solutions dominate by the time the project even starts. The end result? Requirements are written to fit the solution picked, and the real problem remains unarticulated.
Dynamics CRM Ecommerce Integration requires good discovery to document what that one person or department stops struggling with when the project is complete. It is a lot harder to write that one sentence than most realize. Before comparing their work with others, different departments should try writing that sentence individually. Focusing on what different departments stop struggling with helps pinpoint assumptions.
This was discovered during one integration when departments realized a customer’s billing address was described and captured differently in both systems. Each team had lived with the discrepancy so long they had stopped noticing it. The integration requirements they wrote assumed the conflict had been resolved. It had not.
You should collect examples of the current process before mapping an ideal one. When interviewing, ask subject matter experts to walk you through a completed task and explain each step including the workarounds. The process they describe will usually include points where they asked other people for help. This description will often capture requirements that would not be found in a typical requirements document.
A short, jargon-free problem statement is more valuable than a lengthy backlog. Having an idea of what the problem actually was makes it easier to balance the backlog and evaluate solution evidence against the original problem. This is especially important when developing a product to satisfy user needs; the backlog tends to fill with attractive, yet unrelated requests.
Talk to People Closest to the Work
There are types of interviews that yield a ton of notes and very little useful information. You may get someone to describe their ideal state or their day to day job including all the things they would like to have in the current system. This type of information isn’t particularly helpful.
A better alternative is to have someone describe a task they did last week including the parts they had trouble with. This person may also have document artifacts or case work records that capture the information. Often times the workaround described and the case work record do not match. These often include rough notes explaining how decisions were actually made and not how the case team would like them to be made. These artifacts show the actual process and help create a baseline for the solutions.
Interviewing for Dynamics CRM ecommerce integration uncovers something important, like a duplicate contact record, the cause of a confusing service interaction. A situation where a person accessed information that they should not. These are not failure examples. They are design requirements.
Interview the CRM administrator and the commerce team. Then interview the person who is most adamantly opposed to the original idea. Their objections contain the most pertinent design constraints.
What Actually Happens
A workflow map is not a vision document. It does not show how things should be. It shows how things are, including the manual steps, the email threads that are used to transport data, and the person that catches the errors.
Show how data flows through each system. Show where data is created and where it is modified. Show where it is copied from one place to another. For each handoff, show who does the sending and who does the receiving. For each handoff, define what 'complete' means.
Then follow the actual path of a data flow and ask if the receiving team trusts the data they receive. If not, find out how the receiving team corrects the data, and what that process costs. That process is a good candidate for requirement for dynamics CRM ecommerce integration.
Repeating manual entry isn’t automatically an automation opportunity. You have to investigate the reason it occurs. Sometimes it’s a missing process rather than technology, and an integration built to solve the wrong problem creates a new set of issues.
Ask One Narrow Question to Test
A pilot should not try to do everything. This is the quickest way to build a brittle, overly complex solution.
The most useful pilots ask one narrow question. Is it possible to reliably match identities across the two systems and thus remove the need for the current manual effort? Define what would count as a successful response. Define what result would make you reconsider the idea. Test an ordinary case and an atypical, difficult case, perhaps the same contact appearing with two different email addresses across the two systems.
If the idea warrants further consideration, it means the pilot has been successful. If the idea fails because of unanticipated constraints, the pilot is also successful.
Surface the Trade-offs
Planning is an exercise in saying no. Is this decision communicated clearly, or does it show up as a scope problem later on?
Not all features that are valuable are necessary for the first release. A feature that increases development complexity and requires extensive testing carries a burden for the long term that must be evaluated against its value. This evaluation should be done with the people who will be supporting and maintaining the feature, not the people who are just excited about having it.
Write the decision down. A postponed feature is not a rejected idea. It is a commitment to gather enough evidence to justify spending the budget. Having a documented reason why something was deferred eases the planning burden for the next cycle.
Keep Feedback in the Plan
The feedback mechanisms that are designed to catch the assumptions a team makes after working on something for 3 months often get treated as optional.
Feedback is only valuable when it causes something to change. Invite it from the people that were not part of the design process. Their feedback is often critical of assumptions that a team learned to overlook. Look at what people say and compare it to the data. If there is a discrepancy, figure out why.
Turn Discovery into Delivery
A discovery process that does not include the transformation of knowledge into something people can touch, feel and understand is not complete. The end of the discovery process should include something that looks like a delivery brief.
Briefs should summarize the decisions made, assumptions, owners, risks, and next steps including experiments. Without this information, the build team ends up answering a lot of the same questions the integration team should have addressed.
Frequently Asked Questions
Why is discovery so important for dynamics CRM ecommerce integration projects?
Discovery defines what gets built. Too often, teams build integrations that function technically but don't address workflow issues. Some project deliverables may be on time, but many miss the mark because the team is writing requirements based off an assumed solution and not a described business problem. Requirements should be defined by business challenges, not solutions. This guides the creation of integrations that users will actually find useful.
How long should the discovery process be for CRM and ecommerce integrations?
The process generally takes 2-4 weeks for a moderate-sized project integration. The length of the process may be more indicative of the complexity of the systems. The useful measure is whether the process defined a written problem statement, a workflow map of the current state, and prioritized, understood requirements in a way that 2 stakeholders would agree. If these things are defined, then the process is probably adequate.
What should stakeholder interviews cover for a CRM ecommerce integration project?
Interviews should be conducted in an exploratory manner, and stakeholders should be asked to diagram or otherwise represent recent examples of the current process, workarounds, and the interruptions and requests for information that they experience. To aid in their recall, diagrams, workarounds, work requests, and other process artifacts may be provided. Asking stakeholders to describe work they’d stop doing if the integration worked as designed yields a more useful and detailed response than describing features they’d like the integration to provide.
What is a good scope for a first pilot of dynamics crm ecommerce integration?
A good first pilot should test one assumption about integration with a limited audience and define a clear stopping criteria. The pilot should cover identity matching, and the integration of the core fields that matter most to the audience. It should cover at least one difficult edge case, and be focused on determining if the overall approach to the integration is valid, not building a production ready system.
How should trade-off decisions be documented during discovery?
Specify what was considered, what was deferred, and the rationale for each. Provide the name of the person who made the decision, and the cases under which the team agreed to revisit the decision. Though a record should not be long and complicated, it helps to prevent decisions from coming back as unsolved conflicts.
How do you create a useful delivery brief from the discovery findings?
The delivery brief should contain the current-state workflow map, the agreed problem statement, the prioritized requirements, the assumptions each requirement is made on, the named owner for each component to be built, the identified risks, the first piece of production with the named person to verify it, and the definition of done. It should be written in a way that a new employee can read it to understand why something is being built and what is considered a successful outcome.



