The challenges are many with this project, but that is not a point of discouragement. More of an opportunity.
My friend and the first GM that I worked for in car sales in 1994, is a man named John U. John is retired from automotive, but car guys never actually leave the business. John has become the Business Development Manager for an auto group with two used dealerships in Ventura, and Thousand Oaks. Both car lots use Dealer Center as their CRM, and John approached me needing help with email campaigns through the CRM’s bulk email utility. The largest challenge is that Dealer Center does not segment permissions in such a way that sensitive financial data is fenced off from the admin portal. Naturally Admin privileges are required to create templates, and campaigns. So, I have little access to the actual CRM system.
The first challenge was deciding how to create HTML email templates. Platforms like Canva, or photoshop were a complete overkill. The CRM has an editor; however, it’s complicated and makes consistency difficult to maintain. Especially when I have very limited access. There is no way that I would be able to get sales staff on board with learning any of these complicated platforms. So, a web form that generates the templates was decided upon. When I say decided upon, I mean that nobody cared how it worked, just that it worked. So began the perceived simple task that turned into a full blown project.
Entering values dynamically updates the template, and then the HTML that is generated is copy and pasted into the CRM editor. Simple enough, no magic here. Until we needed more templates. Currently there is a version that displays 4 cars, and a generic promotional template that is used for holiday, and special, promotions. Still relively straight forward. There is no back end, it’s all stored in flat files. Completely portable and requires very little resources on my servers. I even wrote an application that ingests images,scales them, and converts them to webP. The car ads use offsite images pulled from the dealer websites, and I had no intention of serving images from my server; however, the random promotional images proved easiest to serve from my host. So, the smaller the better.
Currently, the application allows previews of templates for proof reading purposes, previews of the resized images, and the ability to move them to the live directory. Links can be then copied for use in the form. As I became more involved, I decided hat we needed tracking independent of the CRM. I currently use Matomo analytic server to gain insight into the effectiveness of the campaigns. Eventually, the portal app will have a dashboard that shows the different interactions with the email. It’s just a pain point as Matomo doesn’t allow custom dashboards to be viewed by unauthenticated users. I can serve the dashboard, but only in it’s entirety. I cannot limit any interaction or restrict, or customize, views. Matomo has an API, but I was hopeful that I could avoid that level of development. It is possible to embed specific views through the use of iframe tags, but again they are whatever the default view is. As an example, if the view shows visitors and you would prefer actions … you are stuck with visitors. Hopefully the API offers more fine grained controls.
Some other directions that this may be heading include, serving an iframe of a page on my web server. The template could potentially point to a specific URL that I modify weekly. This would mean that I need even less access to the CRM, but it requires a fall back consideration. A dead man’s switch if you will. What happens if my machine was unreachable, or I was unable to make the modifications? Automating the ads is possible, but that requires a lot of engineering and development. We aren’t there yet.
Another thought was that I have reinvented the idea of a rudimentary ad server. Perhaps it would make for sense to look at using the Revive self hosted ad server platform? Use a tool that was designed for, and continues to be maintained to, do this job. Definitely on the road map for future directions.