Mobile GIS & LBS
Matt holds an MSc in Geography and GIS. He has been working with clients solving problems with GIS for over 17 years. Matt founded WebMapSolutions whose mission is to put innovative, intuitive GIS driven applications into the hands of new and existing users.
What we learned from a failed ArcGIS Implementation
June 18th, 2015 by Matt Sheehan
We pride ourselves on honesty. To not just talk UP about our working world, but discuss the DOWN. Our (my) biggest lessons in life were learned from failure .. not success.
Its time to talk about a WebMapSolutions hiccup. And what we learned from a failed ArcGIS implementation.
We spent the first part of 2015 working with a private company who had no experience with GIS. They had a forward thinking senior staff member who quickly grasped the possibilities GIS might bring to their company. They enlisted WebMapSolutions to help evolve that vision.
4 Step ArcGIS Implementation
Let’s start with some context. We take a 4 step approach to ArcGIS implementations:
1. Planning – defining the problem, and putting in place an implementation solution plan.
2. Data – Finding holes, cleaning, developing schemas, ultimately avoiding “garbage in means garbage out”
3. Platform – Setting up and configuring ArcGIS Online.
4. Maps and Apps – Providing data access a GIS tools for all across the organization.
We followed this tried and tested approach with this client.
What we learned from a failed ArcGIS Implementation
So when we look back on this particular client project, and reflect, what did we learn?
Defining clearly the ArcGIS Implementation Solution Plan
Your ArcGIS implementation solution plan is your project blueprint. It should form the base for any and all projects. It is in this phase that you need the client to explain the problem they think ArcGIS can solve. And for you as an ArcGIS expert to dig deeper, using non-GIS language, to evolve a solution plan based on this problem. Your interactions with your client at this stage are part discovery, part education.
It was during this first phase, with this client, that we recognized our first difficulty:
Challenge 1 – Defining the vision.
The client had difficulty articulating their problem and vision. They had difficulty speaking in non-industry terms, which our staff found hard to translate. The decision was finally taken to put the ArcGIS foundations in place first, based on our 4 step approach, and work hard to understanding better the vision as we moved forward.
Lesson learned 1: Do not move from the planning phase until you fully and COMPLETELY UNDERSTAND the problem/vision. In our case our final documented plan, though thorough and divided into logical phases, was incomplete because the full translation of need to solution had not been made.
Data is at the core of all ArcGIS implementations
Too often data is overlooked. Most clients are focused on the endpoint; usually maps and apps. Data preparation and publishing can be a time consuming process. The complexities of data preparation have not gone away, even though easier ways to publish data from spreadsheets, as an example, are now (finally) a reality. This particular client had some non-trivial data needs. They had no data initially to publish in ArcGIS, which meant first data exploration and discovery. Second, their company data sat in a relational database. A solution was needed to pull this data from their internal system and publish/update it in ArcGIS.
Challenge 2 – Data was discussed at every client meeting. The client never fully understood the data discovery and preparation process.
In part the challenge we faced here was education. The client did not truly understand the effort and expertise required to build a solid base of data. They were confused by the different options, public v private, external v internal, geodata v relational data.
Lesson learned 2: Spend as much time in the data phase as possible. Make sure the client understands and are on-board with the work needed to put the required data in place.
ArcGIS just got too Darn Easy
Esri have made setting up and configuring ArcGIS Online easy. We often encourage our clients to drive this process forward themselves with our support if they need help. This particular client had some experience with the free ArcGIS Online account, but left it to us to put in place their full subscription (even asking us to call Esri to resolve a credit card issue). This saved them time, so we were very happy to help. Once they had the account they started exploring, soon realising how easy it is to publish data from spreadsheets and set up web maps with this data. This led to two quite different platform questions:
Challenge 3 – Confusion over low cost BI solutions and ArcGIS.
Out of the blue we were asked to investigate Domo; a low cost BI solution (one of the owners had been told this would provide all they needed, when out golfing one day). There remains in the marketplace confusion over GIS. Since it has been such a niche technology, the location based analytics it brings is still not well understood. Also the fact that no BI platform (at least that we are aware of) provide GIS integration is also not understood. This question was quite legitimate, and again through conversation we educated the client as to the value of GIS and today’s realities vis-a-vis current BI platforms.
Challenge 4 – ArcGIS “I can do so much myself now, why do I need GIS experts?”
This second issue was significant. Actually the key factor in why our ArcGIS implementation with this client failed. On the surface ArcGIS seems so easy. Why build schema’s, find and clean data when it is so simple to simply publish from a spreadsheet; 5 minutes and we have a web map with a bunch of layers ready to go. Questions began to surface on what we were actually doing, particularly around data preparation. Our meetings increasingly became unproductive education/justification sessions, with questions like “Be honest, how many hours are you actually putting in here?”
Lesson learned 3: Education is ongoing. That is a given. Its a key area of value we bring to all our clients. We worked hard with this particular client around education. The fact we had to work so hard here should have raised some red flags on our end. In hindsight we should have seen this as an area of concern, maybe requesting more or different technical staff on their end to be involved.
ArcGIS Maps & Apps – Finding that “Pot of Gold”
With a plan in place, data prepared and published, and ArcGIS Online set up and configured, we move to the exciting part of any ArcGIS implementation; the maps and apps. This client described the “pot of gold” they thought ArcGIS could help them discover. Giving them new insight into their business and moving them ahead of their competition. We made a decision to publish a subset of their relational data to ArcGIS Online (they really wanted a live feed, but we wanted to demonstrate what their data looked like in a live ArcGIS application by first using a static dataset). Their pot of gold idea was quite unique and required a custom approach, which led to another question:
Challenge 5 – “Why can’t we just configure an app that fits our need and go?”
In a perfect world that would be true. We do not fully agree with Esri on their promotion of configure first. By default, we always look across the Esri maps and apps landscape for solutions which fit our clients needs. But there are many scenarios where a custom solution is required. That can be confusing to clients. This is not the first time we have heard: “We’ve spent all this money on our ArcGIS subscription and now you are telling me we have to spend more on a custom solution”. Now this can be serious challenge.
Lesson learned 4: Upfront, if you believe a custom solution will be needed, emphasize that fact, and repeat it often.
Building trusting Relationships
We really enjoyed working with this client. They were applying ArcGIS in a commercial way, very nontraditional. We believe this type of implementation is the real opportunity and future of ArcGIS. But there are unique challenges, as we found. Where this implementation ultimately fell down was around understanding and trust. We worked very hard on putting in place a solid plan and placed much emphasis in ongoing education. But we failed to build a trusting relationship. We parted on good terms, and wish this client all future success.
As I started this blog post, as a company we learned much from this implementation. We suggest you too avoid shunning failure; be stronger and better for the knowledge you gleaned from the experience.
7 Responses to “What we learned from a failed ArcGIS Implementation”