In my last article, I outlined how we in Snowflake Support use contextual data about where our customers get stuck to improve the overall product experience. Now I’ll take you through how your organization can also implement these important feedback loops from support to product enhancements, to your company’s—and your user’s—benefit. 

Customers don’t wake up in the morning and decide they’d love to spend time with a Support team. Knowing this, we are mindful of removing reasons why customers contact us, and reducing the effort required to get to a solution. Here are three strategies for implementing such a program.

Unify around a common cause

Pick a focal point for your Customer Experience Analysis (CXA) efforts. Ours is a list of what we call Hot-Spots, which is any recurring pain point or time-intensive case trend sourced from our extensive support interaction data. Perhaps there’s a recurring type of case in a single area of our product. Or, it could be an emerging high volume of cases popping up immediately after a feature release. Or, any case requiring greater-than-usual effort from support or tasks that require contact with our Support team.

Regardless of how we slice and dice the data, each Hot-Spot candidate is well documented in a Jira issue with all the context we can find, and includes detail from support cases, trends, and support engineer feedback. 

Get cross-functional buy-in

CX analysts communicate each new Hot-Spot to the relevant parties in recurring syncs to align on familiarity and prioritization. Some Hot-Spots require further context that only our Product or Engineering peers can provide. Either way, we have a focused set of support-driven issues to rally around. We’ve gotten signals that this approach is working, too. After CXA shared Hot-Spots relating to his area of the Snowflake product, one senior engineering leader said, “Fantastic job emerging the data; this is an amazing example of how our teams should collaborate.” 

To make buy-in simpler, CXA meets our internal customers where they are. We curate Hot-Spot with Product and Engineering prioritization in mind. We curated the process in such a way that neither the Product or Engineering teams need to add any level of complexity with a new project management tool since the Hot-Spot program works within their already established, respective workflows. Hot-Spots live in Jira and by this point have already been socialized and scrutinized—most are now ready to be prioritized in quarterly product roadmaps and planning cycles. 

Operationally, a core team of CX analysts and Support and Engineering executives meet regularly to maintain progress, prioritization, and accountability—and celebrate successes!

In 2022, CXA program efforts extinguished Hot-Spots for our customers. Here are a few highlights:

• God monitors: Teams could be brought to a halt by a legacy, internal resource limit. Before we flagged this to our Engineering team, 80% of the cases caused by this limit were Severity 1 and 2, and required assistance from the Support team. Customers couldn’t solve it themselves, which was a suboptimal experience—but this has since been rectified.
• PrivateLink enablement: AWS PrivateLink is an AWS service for creating private VPC endpoints that allow direct, secure connectivity between your AWS VPCs and the Snowflake VPC without traversing the public Internet. Before our Hot-Spots program highlighted this recurring issue, it required a support case to enable PrivateLink. Now PrivateLink enablement can be implemented by end users. Today, more than 85% of these tasks are self-served with a well-documented Snowflake function.
• Automating empty credential pools: We discovered multiple cases resulting from a lack of credentials during the creation of storage integrations, and the potential for this to cause future support cases. Snowflake-side automation enhancements took us to a better future without this potential disruption.

Defining how CXA benefits both customers and employees

Who knows customer pain points better than our team members who’ve solved thousands of support cases? No one, of course. Tenured support engineers have vast amounts of technical, organizational, and customer knowledge gathered over years of experience. Snowflake takes great pride in offering impactful career growth opportunities for those who demonstrate excellence and are ready to widen their sphere of impact. To that end, all of our CX analysts are former CSEs who bring both technical and soft skills to the job, including: 

1. Data analysis savvy: We use all our support data sources and tools to identify patterns and trends, and challenge ourselves to find the customer’s voice in everything.
2. Attention to detail: Seemingly small details can indicate larger issues. We watch for the slow or quiet Hot-Spot arrivals, too.
3. Customer focus: Like other business leaders, we believe that the customer experience “is at the heart of creating intentional futures.” Our CX analysts have a demonstrated record of customer empathy and weave CX into everything they produce.
4. Solid communication skills: We work through other teams to prioritize and resolve Hot-Spots, so it’s crucial to be a confident writer and speaker.

If you’re inclined to start a CX-focused initiative, you probably already know more than you think you do and could kick-start one quickly in your organization. If you can’t already name half a dozen issues that are Hot-Spots for your support organization, ask some basic questions of your support data using these sample question/answer approaches to find some low-hanging fruit:

• For case cohorts that show significant spikes to KPIs, such as new case creation rates, median resolution time, or case severity distribution, week-to-week:

Q: What are customers saying in Customer Satisfaction (CSAT) transactional surveys for that same cohort of cases? How about Customer Effort Scores (CES)?
A: Maybe customers are already telling you directly about a problem. Or, you can connect with your support case handlers who have a read on what’s trending.

Q: What else happened at or around the time of the spike?
A: Maybe the spike coincides with a new release or updated documentation; taking inventory of other events may help you identify a gap you can plug to avoid a future spike and resulting customer impact. 

• Within your highest consistent case-drivers per week, ask: 

Q: Are there case types here that *require* your customers to contact you?
A: Strongly consider self-solving options.

Q: If there is case high volume and high throughput (low median TTR)
A: What can you do for this case cluster to drive that median TTR to zero? 

• Find 30 or so of your oldest cases (still open or already solved), or cases that are far outliers compared to your median TTR. Such cases almost always exact a high cost, occupy your most senior technical resources, garner the lowest Customer Effort Score and Satisfaction scores—but they can also teach you a lot.

Q: What attributes do they have in common?
A: If they all pertain to a certain area or feature of your product, ask your supporters if their troubleshooting tools are adequate. If you leverage Swarming/Collaboration, are case owners making use of the collective knowledge available, early and often enough?

Q: Poll the owners of these oldest cases: “What single thing do you think contributed most to this case staying open for so long?”
A: Maybe a dated process, outdated knowledge content, or training needed for a new feature? But be willing to hear something you didn’t already consider!

Put a CXA program to work for your business

Our CXA program is working because we rallied around a core company mandate to “Put Customers First” and proved that our support interaction data contains a wealth of customer sentiment we can act on. We’re proud of our CXA outcomes thus far, and we’ve got the foundation set up to extinguish dozens more Hot-Spots in 2023. We also hope that sharing this approach can inspire other support organizations who want to reap similar value! 

The post <strong>3 Steps to Creating a Support-to-Product Feedback Loop in Your Organization</strong> appeared first on Snowflake.