Last Mile Data Analytics and How to Achieve It
We all know that data on its own has no value. It needs to be productized, packaged, and ultimately delivered to its final destination on-time, in-full, and in the right condition.
As data leaders, you don’t just know that. You feel it because it’s also the problem that leaves you and your team members in complete and utter despair. “Why didn’t we simply choose the easy life? Why didn’t we all have the sense to be product managers?!” your team cries as they shovel their way out of another ugly spreadsheet.
Things aren’t quite so bad, because we can not only solve that problem with the right tool, but there’s also another way to frame the problem to help our comrades in business understand just how big this problem is, how costly it is to the business, and how fixing it benefits all of us.
When you think about it, delivering the value of data is just like parcel delivery, otherwise known as last-mile logistics.
We’re not the first team to make this connection, but we do have our own take on it. So stick with us for a minute.
What Do Data Teams and Logisticians Have in Common (Besides Everything)?
What does your data team’s dream of using their expertise and skills have in common with a logistics director?
Well first, you’re both asked to use spreadsheets to do things they have no business doing. Your team is spending hours, days, and weeks trying to sort out some thrice-copied spreadsheet with a vague title in Google Drive in order to derive value from it that they could have created with beautiful code in minutes.
Logistics teams are in the same sort of hell. Imagine trying to keep track of where thousands of shipments are at any given time with capacity crunches, inclement weather, border delays, and more…in a spreadsheet. (Okay, maybe data science isn’t so stressful after all.)
The commonalities don’t end there. In fact, we think they peak when both our products reach the last mile.
The last mile in logistics is the most expensive part of logistics. Why? There are a lot of reasons. At the crux of it, though, is that when you have more parcels out for delivery, your logistics program becomes more resource intensive.
With every added parcel, logisticians’ ability to achieve 100% delivery gets further away as does their ability to focus on program improvements that earn them their bonus and add value to their job.
See the connection? It’s the same infuriating problem that leads to burnout for data teams, prevents you from adding value to your business, and costs everyone time and money (and precious sanity points).
The good news for logisticians: there’s a technology solution for last-mile logistics. It started with milestone data from folks like FedEx and has culminated in aggregators like Parcel Perform.
But there is no FedEx for data. Or at least there wasn’t until we built one.
What’s Previously Prevented Last Mile Data Analytics?
Moving towards last-mile data analytics requires acknowledging and more importantly, articulating a problem that we have and will continue to have for the foreseeable future.
Here’s the problem: Business teams that want data don’t feel they can get access to it or work with it. At the same time, our data teams want to provide that data to business teams, but they don’t operate with the same tools (usually spreadsheets). The business team creates their own work, usually in a spreadsheet, and the data team then spends their resource time translating a spreadsheet into something remotely usable. And the process repeats itself over and over again.
The result: business teams continue to create Jira tickets with a rough estimation of the data they need based on their own locally-saved spreadsheet.
And your teams work through tickets until they burn out or quit or really do consider becoming product managers.
Meanwhile, data leaders are trying to justify their team's existence to their executive stakeholders rather than showing off the true ROI of a strong data program.
If we go back to our logistics analogy, the similarities are obvious.
Logistics is all about collaboration between teams and stakeholder management. If you’re a logistician, your boss expects you to strike the unachievable balance between lowering costs and maintaining quality. Meanwhile, your sales team sold products that you don’t have yet. And then there are your logistics service providers, who change the goalposts and think you won’t notice, all while the contracted motor carriers all do their sweet own thing. It’s chaos.
What is the Answer, if Not More Business Intelligence Tools?
The good news for logisticians is that there have been solutions for logistics, including last-mile deliveries, for the better part of a decade. They’re imperfect but they address the core problem and they allow less firefighting and more decision-making.
For last-mile data analytics, the same is true. We’ve worked around this problem with imperfect tools like business intelligence dashboards, which solve an entirely different problem but definitely not the one we’ve just described.
For most of us, a combination of Python notebooks, spreadsheets, and ad hoc applications is the so-called solution. But these are workarounds, and they’re not great because they’re not getting to the heart of your problem, which is the inability to collaborate effectively while using the tools and skills most available to each team.
What you need is a solution fit for purpose and focused on solving the problem, not a workaround that perpetuates it.
The principles behind the tools we need even share a lot of similarities. In the data context, we need a solution that is:
Robust: You need a solution that can handle large data tables and complicated models as well as simple as sorting through weekly revenue. In our logistics analogy, this means a tool that can handle the delivery of a crate of tennis balls as well as a solution that can support the delivery of cell and gene therapies, which are worth millions of dollars a shipment and have much stricter shipping conditions that must meet regulatory requirements.
Plastic: You need to be able to deliver data anywhere to meet your teams where they are, which means being able to work across contexts. Without this principle embedded, the tool won’t calm the chaos. The same is true in logistics: if the tool isn’t global and multi-modal, it might as well just be a spreadsheet.
Fast: You need to build analyses quickly because your teams have a never-ending list of tickets to work through and the due date for everything is ASAP. But you also want to deliver these quickly because you don’t want to spend time on this work because it’s not value-added work. The correlation with logistics? It’s obvious. (And it’s all Amazon’s fault.)
Remember how we said we had a solution? Now it’s time to dive in. And if you’re skimming this article, here’s where the value is.
Delivering Data On Time and In Full to the End User — Without More Chaos
Spreadsheets won’t die; BI tools aren’t the answer; and if we want to keep talent in data science, we can’t continue with the workarounds.
So Arcwise created a tool that will help you expand the impact of your data products that is robust, plastic, and fast so that it won’t cause more chaos in your data and tooling.
And it’s not (NOT) another dashboard. Seriously.
We’ve solved for the real problem (a lack of space for collaboration), so business teams can stay in their spreadsheets, and data teams don’t have to compensate for anyone else’s lack of Python and SQL skills or compromise your data architecture.
Arcwise is a tool that lets you stay in your current Google Sheets and allows your team to run any analysis, at any scale.
How? Our tool translates your spreadsheets into SQL, so your data team doesn’t have to wrangle another spreadsheet, ever again. Imagine the possibilities of getting data to the right place, in the right condition, and without all the time-consuming busy work. It’s data scientists skipping through meadows of wildflowers, thinking about how they’ll stay in the field forever because they can use their skills to add so much value.
And as for logisticians? They could be so lucky. We really, really feel for them. But then, there’s always product management.