Give a sales person data and they’ll have insight for one day but teach a sales person to self-serve and…
At Cox Automotive, the idea of self-serve reporting isn’t new. For many years we have had a reporting portal available where users can log in and download excel sheets of data. This, along with a support team running ad hoc data requests, has served Cox Automotive well.
However, we have always wanted to provide more for our users. Downloading excel sheets leaves a lot of work to be done to find and present data in a format that is helpful to them. Dashboards provided us with a mechanism to create appealing visualisations for our users and to give them the flexibility to query them as the saw fit. We didn’t know it then, but this was our first step to truly empowering our users. This is the story of our journey to self-service and beyond. From running to stumbling to running again…
The good old days
5 years ago, sales teams at Cox Automotive requested data directly from analysts. Each request was interpreted by the analyst, who manually queried the database. Data requests could take up to 48 hours to turn around, and longer during busy periods. And the output wasn’t anything fancy. There would be calculations sat behind some of the data, invisible to the end user. Some were commonly used, others bespoke to the analysis but essentially, we were extracting data from a database in the form of a spreadsheet.
Once the data was exported, the query would either be saved for future use or deleted. The users would then go away and produce charts and whatever additional calculations their customers were comfortable seeing.
This was time consuming for the analyst and the user. We were often duplicating work but for a long time this was the way it was done.
Around this time, dashboard reporting was becoming more accessible and we seized this opportunity to try and change and improve the way we worked. We knew there was an opportunity for this to change.
Empowering our users
Introducing a dashboard-based, self-service platform was revolutionary. It allowed users to access the data when they needed it. Dashboards were available 24/7 and the graphs and charts created by analysts were designed to present the data in a consistent and branded way. No more waiting for an analyst to get around to your request, no more senior stakeholder taking your place in the queue. Self-service dashboards liberated both analysts and end users.
The golden age of Qlik
The first platform we used was QlikView, at the time the leading provider of BI tools in the industry. It scored points immediately with users. It was extremely quick at displaying data and served internal Cox Automotive users well with a few well thought out dashboards. However, as time went on, requests for obscure data sources to be added or data twisted in different ways started to feed through.
We tried to include many of these requests with the current suite of reports until it became apparent that data would once again need to be extracted and manipulated in excel by the user…meaning that the user would be once again spending time managing data in excel.
No such thing as a free lunch
QlikView’s licencing model was very expensive for this potential use case. Charges were placed on dashboard users so we couldn’t create disposable dashboards, and handing out desktop developer licences was out of the question. This resulted in a sales dashboard containing data sliced in so many ways that it was rapidly becoming very hard to manage.
What we needed was a reporting platform that not only had the formal dashboards of QlikView (i.e. constructed after consultation to cover the core needs of the business area they served), but also give the user the option to make their own dashboards. This is when we discovered Tableau.
New sheriff in town
Tableau appealed immediately. With a far less restrictive licencing model we would be able to create smaller bespoke dashboards, better suited for the quick responses the teams required. It was simple to use and performed well as a dashboard creation tool. It was able to connect to a large number of data sources (including big data connectors) which was of increasing importance.
Most of the hard work had been done over the previous year engaging users and promoting the benefits of self-serve reporting. People were enjoying the freedom this 24/7 access to data gave them, so now we could take the next step and move the users from using pre-built dashboards and reports towards building their own.
Tableau Web Edit
Web Edit in Tableau enables the user to create their own server-based dashboards from any data sets available to them. These dashboards can then be made visible to any user they choose. They enjoy the same benefits as analyst-built reports, such as auto updates, security privileges and backups and are available 24/7. With this feature we could give the power back to the user and in turn free up our time to spend on other projects.
No need to become a mind reader
There are inherent problems when creating a report for someone else, such as gathering and interpreting their requirements. Some people can articulate clearly what they want and why. Others are vague and require more iterations to get it right. And some people just don’t know what they want!
With Web Edit, we can remove ourselves from interpreting the enigma of the user’s requests and pass this back to the requester, who will know “right” when they see it. But, this can only work if we promote the value of web edit and provide enthusiastic support to any adopter.
Engaging with our users
In the past we used mailshots to promote high profile Tableau dashboards early on. We used the same method to announce the web edit facility. To speed the adoption up, we identified users who showed a keen interest in Tableau (above the normal consumption of pre-made material). I’m pleased to say, we also had people approach us for this facility.
Training users was simple. We created simplified data models for the web edit users to use and a formal training schedule was drafted. For 6 weeks, we held an hourly training session for small groups (no more than 3 people).
The format has been a great success and the engagement from users has been particularly pleasing to see.
Since the introduction of Web Edit, we have completed two sets of training sessions. One to our Sales team and another to our Buyer Services team. Both teams are heavy Tableau users and frequent requesters to our internal request system. We now see a smaller number of requests coming from both these teams.
Still room for improvement
That isn’t to say it’s all been plain sailing. Tableau Web Edit has limited functionality compared to the desktop version. Giving users the ability to create their own reports comes with a cost. Calculations can be created that don’t do what the users intended and inappropriate visuals will still limit a dashboard’s interpretability. However, any changes in Web Edit are traceable. We know who has made the visual and can give guidance on best practice. We can see the rogue calculation and help the user fix it.
Tableau regularly release updates and have an active forum for feedback and assistance, which has resulted in a great product that continues to impress on a visual and analytical level. With recent updates the Web Edit functionality has been improved and is now a feature rich experience that remains easy to pick up.