I recently went through a three-year process of having a custom car built. It was a long, expensive, and exhausting (no pun intended!) effort. During this time, I noticed many similarities between what I went through and what many organizations go through when it comes to their digital analytics implementations. So, I thought I would share my observations using building custom cars as an analogy for building analytics implementations.
I used to own a 1958 Corvette. I love the first generation Corvette and always wanted one. Here is a picture of my family and me in my first Corvette (notice the license plate!):
Unfortunately, this car broke down all the time. I spent more time towing it than driving it! This reminds me of the countless organizations that have bad digital analytics implementations. My car spent hours in the shop, and at the end of the day, I didn’t get much value from it, just like organizations that derive little value from faulty analytics implementations.
To make matters worse, while I love cars, I have never had the time to learn how to work on them myself. This meant that any time something went wrong, I had to pay someone else to fix it, which just added to the overall cost. In the analytics world, I have seen that many organizations don’t have people on staff who really know their digital analytics tool well enough to be able to fix issues or improve their implementation. That forces them to bring in outside consultants in the same way that I leveraged mechanics (and those analytics consultants cost more per hour than mechanics!).
After a few years of sinking money into my ‘58 Corvette, I decided to sell it (and when I say “decided,” I mean that my wife forced me to). Little value plus a lot of money was not a recipe for success. The same is true with digital analytics implementations. Eventually, organizations tire of throwing good money after bad on implementations that aren’t producing value.
Unfortunately, you can’t sell your analytics implementation like I did my car. Instead, organizations are faced with the decision of either doing a full re-implementation or moving to a different analytics tool. As I mentioned in a previous post, switching to a new analytics tool is often the wrong choice, since it is simply a band-aid for the real underlying problems. In my car situation, I decided to sell and hope that one day I could find a way to get a similar car that actually worked. But here’s where my analogy diverges from my analytics recommendation: In most cases, I recommend that you keep your existing analytics tool and re-implement it to meet your needs.
So, fast-forward a few years, and I am visiting ObservePoint in Utah and John Pestana shows me his first-generation Corvette that has been completely re-done with a new engine, transmission, and all of the features of a brand new car (shown below)!
This was exactly what I wanted, so I got the name of the shop that built him his car and decided to build something similar. I relate this to an organization seeing a presentation at a conference of an awesome digital analytics implementation and deciding that they are going to hire the same firm to build their implementation. But, at the end of the day, I was looking at a completely custom project, with a lot of moving parts, building something custom that I really didn’t know that much about. What could possibly go wrong?!
So I contracted with the shop John Pestana used, bought a car with the intention of modernizing it, explained what I wanted, and started writing checks. Unfortunately, as luck would have it, the shop John used had some major employee turnover, and the people there weren’t as experienced as those who had done the previous work. I have seen similar situations in which consulting firms have a few superstars who leave and the consultants who remain don’t have the same skills. It wasn’t long before I noticed my project stalling. Not much was getting done and I grew frustrated. After almost a year, my car looked like this!
It was clear that the shop was not getting the job done, just like the times Adobe Analytics customers call to beg me to help fix what their internal teams or an outside firm had done with their implementation. I decided to switch shops and found another one that a friend told me would be much better. I had sunk a lot of time and money into the first shop, but it was time to move on.
The new shop was much better, but they had many projects they were already working on, so I had to wait a while. The best shops, like the best consultants, are often worth the wait, and there’s a reason they have people lining up!
Unfortunately, the new shop informed me that much of the work done by the first shop was done incorrectly. They encountered shoddy bodywork and even found that the front of the car was being held together by a hidden 2×4 piece of wood (likely from a past accident). The first shop had no idea about this! This is similar to what I find when I take over Adobe Analytics implementations from other firms. Organizations spend a lot of money and then get annoyed with me when I tell them that the other firm did things wrong. That requires them to pay again to fix it, which never sits well with clients. But at least I felt like the new shop knew what they were doing and that finally I was in good hands, a feeling that I hope my Adobe Analytics clients feel as well.
Once the new shop started, things went pretty well. There were a lot of key decisions that we had to make together, but in the next year, the car progressed and was eventually completed. Here is the finished product:
So in my case, things eventually turned out ok. I spent way more than I had planned and it took way longer than anticipated, but I eventually got what I wanted. I expect that this is how many organizations feel after an analytics re-implementation as well.
Of course, something broke within the first two weeks of driving it, so it is already back in the shop since I am not knowledgeable enough to fix it myself. Like an implementation, sometimes things will break and cost more money, but if the bulk of the work was done right, everything should run pretty good.
So, as you can see, there are a lot of similarities between building a custom car and building an analytics implementation. But this entire saga got me thinking about something bigger. When people see me in my fun, new car, they love it. I get waves and lots of great comments. People tell me they wish they had a car like mine. So why does 99% of the population drive cars that roll off assembly lines?
The need for multiple efficiencies drives the car industry and it’s what’s driven the analytics industry’s need for Apollo, the first repeatable automated implementation solution. I’ll explore those correlations in my next post…
