How IT and business work together - a business systems case study from SXM Media

Every business faces a growing need to 'fill in the gaps' around existing enterprise applications in the search for greater efficiencies and better customer service. Doing this successfully demands close collaboration between the IT department and business functions — which is easy to say, but often challenging to achieve. Every year, the Business Systems Magic conference brings together practitioners that work along this fault line, preferring to talk in terms of business systems and business technology (BT) rather than IT, to emphasize that they have their feet in both camps. Among the sessions in last week's virtual event, Kim Mestre and Andrea Bromberg from SXM Media spoke about their work to deliver automation to a busy revenue operations organization.

SXM Media is the combined sales organization of the audio entertainment platforms SiriusXM, Pandora, and Stitcher, and also sells advertising for others in the industry such as SoundCloud and NBCUniversal. This makes for a challenging business systems landscape. The business needs are sophisticated and highly digitally driven, while the IT systems are large-scale and complex. Furthermore, in an organization that's highly focused on revenue, it can be hard work persuading people to carve out time to pursue an automation project, as Bromberg, Senior Director of Client Services, explains:

I need to go to my boss and my boss's boss and say, please carve out time. Please can we have some people and take them away from revenue driving activities? Do you launch a campaign over the weekend that's going to bring in x hundred thousand dollars? Or can I use that body and help me work on this automation stuff?

The key to success is identifying the metrics that will demonstrate the worth of the automation project. These KPIs can be anything from revenue saved, overtime numbers, data accuracy or error reduction to client satisfaction. This means the intiative must be data-driven, as Mestre, Senior Process Architect for Intelligent Automation, explains:

When we initially start talking to a team about a project, yeah, we want to hear about what the process is. But we want to know those KPIs because we have to track those, they want to know them. And then those can be reflected in the case studies as well.

Keeping pace with rapid change

Their recipe for success is a combination of this data-driven approach, leadership buy-in, and a lot of communication at all levels. Collaboration between IT and the business function is centered on a Center for Enablement (COE) structure. Mestre is part of the IT teams' COE for intelligent automation, working with the Director of Automation and a developer. Its job is to evangelize intelligent automation across the company, to work closely with both the teams that it is automating processes for and the development team, and to continue building up the digital automation toolkit. On the Client Services side, the counterpart COE consists of Bromberg along with managers from various divisions, both within the in-house organization of around 250 US employees, and in the offshore support teams.

Client Services works closely with the SXM Media sales teams, providing data reports, solving ad tech issues, working on renewals, and so on, across a diverse client base that ranges from local mom-and-pop stores to global brands. The organization has to manage rapid scaling as it grows, digest acquisitions such as last year's Stitcher podcasting platform, and adapt to change in a fast-moving industry. This is the backdrop to the quest for more automation. Bromberg says:

We can sum up the problem that we're trying to solve as, how do we keep up? How do we keep up with the industry? How do we keep up with our changes? How do we keep up with everything that we're trying to do? And not just stay afloat? ... It's not just keeping up, it's how do we stay thriving in this environment?

While product changes can take months or years, there are many other elements in the day-to-day operations that could benefit from standardization and systemization, but which are not always top of mind. She continues:

When you think about a world where we're moving 1,000 miles a minute, we don't have product standardization. We don't have systematization. There's a lot of just Google Docs and people knowing things because they've been here for a while. We don't want that knowledge walking out that door when they do. So how can we make sure we're setting ourselves up for success in the future? And making sure we're keeping all that knowledge within our organization, somewhere easily accessible to all people, and not losing it?

Building on an early success

Previous efforts to bring in automation had failed to deliver because they had run into hidden challenges. There are many different systems across the organization, from enterprise applications such as Oracle and Salesforce to more specialized systems, as well as all the ad tech infrastructure and various internal systems built by the in-house product and engineering teams. This patchwork of systems has a very disparate ownership landscape, each with their own roadmaps and timelines, while the management structure across SXM Media has a lot of different layers and demarcations. Negotiating all of this takes leadership buy-in and a lot of careful communication, which is part of the reason for setting up the COE structure. Mestre says:

A really big part of why our collaboration has been successful is that they have been able to dedicate resources to this program. And so leadership is asking questions, now we're being held responsible. I think that's what we needed. Because if it's just a band-aid solution, then there's probably not going to be a lot of follow-up. Having a structured programme in place is really what has helped us be successful.

One early success for Mestre's team also provided a case study that demonstrated the value of automation. This automated the finance team's Insertion Order (IO) audit process. Once a client's order for an ad campaign has been finalized, the IO is drawn up to detail all the components of the campaign — rates, slots, content assets, number of days and so on. This PDF document goes through a variety of approvals, from legal to pricing and sales, and the final step is an audit carried out by the finance team to validate that everything in the signed-off PDF has been correctly entered in the operations systems by Client Services. That manual process could take from three to five business days. Mestre's team used OCR to read the PDF and extract the data into a database that a bot could then compare against the systems and alert Client Services of any errors. This cut the time taken to complete the audit to around half a day. This was a powerful case study, says Mestre:

We've not only saved tons of hours, but we've expedited the time in which it takes an audit to happen, and therefore for the campaign to actually go live ...

We were able to come to them with very concrete, and what happened to be excellent, numbers to help prove the case study. That we were able to solve for X percent of the insertion orders, we saved the Shared Services finance team X percent of hours.

Finance is "a no-brainer" for automation, she adds, because there are usually very clear-cut rules for what needs to happen, and less scope for the kind of variations that you might get in a creative or marketing process. This is a guiding rule for applying a Robotic Process Automation (RPA) tool like UiPath, she says:

What are the most manual, time-consuming processes with the most clearly defined rules? Then we go automate that.

Getting a quick win is crucial for building trust. As a starting point, Mestre recommends finding something that requires minimal development effort — for example, merging some reports, applying some business rules and then emailing out the results. She explains:

Once we can get something to production, that is when the team can really understand the impact. Because how is a bot doing this? None of that really makes sense, until that email arrives in their inbox, and they didn't have to do any work to get it there.

Importance of communication

Once people see that the outcome frees up their time, it relieves some of the anxiety they may feel that automation will take away their jobs. Bromberg elaborates:

If we can find all those little moments and all those little button pushes and all those little six extra steps to get the report that you need that we can eliminate from their day to day, they're able to uplevel themselves so much in a so much faster way. They're able to really focus on the strategic and revenue generating things instead of the day-to-day button pushing that we sometimes get caught in.

So I think positive sentiment and good PR is not something to sneeze at. I'm having people knock down my door, saying, 'Hey, I'm on this one project, and maybe RPA can help, can I talk to you about it?' And that, to me is such a win.

Getting it right means taking time to talk to people that actually do the work and know the processes, she adds:

This can't be something that comes just from me or my boss of what we want to fix. It's asking the right questions and making sure we have the right people that touch the right pieces of the business, and listening to them and hearing them and vetting and understanding at all.

Communication is important on the IT side, too. Keeping in touch with the product team is essential, to understand what they're working on and whether there are releases in the pipeline that might impact an automation project that's in the works. Collaboration tools play their part too, as Mestre explains:

We rely heavily on Slack and Google Docs so we can all collaborate together ... Those tools are very helpful because we work across four different time zones. So all of America, and we work with India. And so it's almost impossible to get us all on a Zoom at the same time. It happens sometimes. But that's why being able to comment on something in a Google Doc, and then we know that the developer offshore will be able to pick it up when he's back online.

While the COE works on processes that can be used across the organization, there's also a role for 'citizen developers' to automate their own individual work routines, adds Mestre.

If there are individuals that are excited to automate their own things, then we can also supply the software to enable them to do that. And so if you're thinking about maybe you have a process that you run on a daily basis, you can have a bot on your machine like a desktop assistant that you can trigger and run that whenever you want ... The difference is, it's just doing the work that you would do, versus when we're developing something under the corp umbrella, it's automating it, for example, for all of the account managers.

My take

A valuable reminder that, when working with technology, often it's working with people that matters most.