5 considerations for scaling support from customer service experts

When it comes to customer success at high-growth startups, a question we’re constantly asked is, ‘how do I scale support effectively?’ This question tends to emerge once a company has an established set of customers and the infrastructure suddenly feels like it can’t support the needs of the growing customer base. Scaling support is a natural growing pain that forces businesses to mature their operations; they must build programs that balance service with expectation setting with customers, all while designing an efficient and scalable team structure.

Below are insights on five scaling topics from seasoned support leaders in the Primary network: Pooja Radia (Great Jones), Eleanor van den Heuvel (Coinbase), Ronan Conroy (formerly Sailthru, DoubleClick and Adobe), and David Fox (CM Group).

Balancing quality with scale

The quality of support, especially in the early days, is hugely important. As the company scales, the quality of support might look different, perhaps not as high-touch, but it shouldn’t be seen as trading quality for scale, rather the right investment needs to be made to do this effectively.

Pooja acknowledged that high-touch customer service is key as the business acquires its earliest customers, and setting the right foundation from the beginning is important before making any scaling decisions. “You want to be able to achieve and boast a strong NPS and accelerate growth. This might mean a higher-touch service experience, but it doesn’t have to come at a high cost. I recommend first setting up a best-in-class tooling and infrastructure system for support, platforms like Zendesk or Kustomer, then setting the right KPIs for the team... After you have that setup, you can think about bigger scaling decisions like domestic or offshore to reduce cost.”

For Ronan, the first step in scaling support is evaluating a big factor, and addressing a basic question. “How many staff members can you put to work in front of the customer, handling incoming inquiries? You’re dividing a set amount of investment dollars by a number of staff members, which means you’re trying to figure out the optimal salary level.” Ronan warns other support leaders to avoid the trap of throwing more humans at the problem and expect it to be solved. For more technical products and support, he recommends assessing “the inherent trade-offs in scaling quickly by distributing salary investment in more support headcount. You may in fact end up paying a price in terms of the level of hand-holding, ongoing training, needing to intervene in escalations, and perhaps if situations are handled less ideally, you may end up with some expensive fall-outs.”

At the most basic level, Ronan says you have to evaluate why the work is coming in and what the teams (beyond support) have to go through to get the work addressed. Questions he recommends asking: “Are there gaps in the product that lead to issues, or confusing workflows that lend themselves to customer errors or confusion? Are there bugs that create issues customers need your help with? Are there recurring or administrative tasks customers cannot complete on their own and rely on your team to complete them?” Evaluating what those gaps are and working back from there will set the path forward.

Similarly, Eleanor recommends thinking about support through the customer’s lens when evaluating the metrics to consider in the tradeoffs. “It is always worth prioritizing and working backward from success metrics based on what truly resonates with your customers. Most customers — most people — want their issues handled with a minimum level of effort. This means first contact resolution (FCR) is the primary metric for most scaled support scenarios. Other metrics can support and inform this primary metric (transfer rate from generalist to specialist, for example), but regardless of how support is tiered, using FCR as a north star can ensure the function is vectoring to a customer-centric place.”

Deflecting tickets? Measure, measure, measure! And build a knowledge base.

Ticket deflection can be broken down into three categories: elimination, automation, and optimization, says Eleanor. “In elimination, we look to ensure there is a firm feedback loop with upstream teams (e.g. Product or Engineering) to ensure they’re prioritizing top customer pain points (bugs, poor product flows, feature requests). For automation, these days this is largely AI-driven. Customers, in many cases, want a self-serve model and if you can provide them a path to do so via automation, this is a good path forward. The last area of opportunity, if you can’t eliminate or automate, is to really optimize the customer experience to ensure it’s fast and easy for the customer. The best approach is an interconnected one between these three areas, which each inform the other.”

David says the key to programmatically deflecting tickets is building internal and external knowledge applications along with customer-facing product configuration tools. When it comes to rolling out self-serve ticket deflection specifically, he says “the number one rule is crawl, walk, run. Or as I sometimes say, deploy quickly, learn, adjust, repeat.”

In Ronan’s experience, increasingly more customers want to have the answers without logging a ticket or making a call, or waiting for someone to investigate and get back to them. To solve this, “build a knowledge base with frequently asked questions and information on how customers can solve problems on their own to help deflect ticket volume.” The challenge can be pulling this knowledge out of the support team’s heads, as not every support person feels comfortable writing this content, but he recommends investing in incentive programs for support members to write content and investing in tools that can convert ticket content into article drafts.

Dedicate the time in setting up the infrastructure for issue type and accurate data tagging so that you can tackle the tickets that are driving the majority of tickets, and assess whether they can be or were solved in one touch, Pooja recommends. “Start small by building a knowledge base and content for the issue types you’ve identified as high volume and easily resolved to enable customers to self-serve. Then launch a help center tied to your service desk that basically forces customers to find their answer before putting in a ticket.”

The biggest driver to set up a central knowledge base, and understand what needs to be eliminated, automated, or optimized, is to measure everything. “Know your baseline, and rigorously measure the impact of these initiatives by looking at the number of tickets tagged with a specific issue type in a particular time period. Know when you need to kick in product marketing to increase adoption of these self-service features.”

Speaking of knowledge bases, some tool recommendations:

David has used Salesforce, Zendesk, Freshdesk, and lesser-known CRMs. He advises “when you purchase, there are a ton of considerations. My top three would be ease of use, broad integration capability, and robust reporting.”

Zendesk is a favorite and one of the most robust tools in the market in terms of how it integrates with other tools. Guru is a recommended knowledge base tool that integrates with Slack and Chrome features. For 24/7 support, Ronan has usually relied on a phone system that allows toll-free inbound calls connected to the ticketing system, and that easily connects to a paging solution like PagerDuty to alert on-duty staff when necessary.

For consumer-facing support, Eleanor says the primary mechanism is machine learning/AI, and messaging platforms have come a long way and are very robust. She also stresses that UI is imperative to the support experience, and she has seen companies under-invest in tooling for support agents, not bothering to customize tools like Zendesk and Salesforce to integrate them with other systems that agents use to address customer issues. “Optimize the support tooling and everyone can work in support of the most important thing: the customer.”

How to ensure customer data is driving product decision making:

Unsurprisingly, capturing customer data (and questions!) are invaluable to the product teams. It’s imperative that the product team is seeing data, and David recommends giving the product team the data even if they aren’t asking for it. He recommends consistent meetings with product and engineering teams to ensure customer insights are being heard and suggests a business analyst own actioning and project managing that data.

A Voice of the Customer (VoC) report or dashboard is a commonly accepted mechanism that helps that feedback loop to the other internal teams, specifically product. Eleanor says “product should be viewed as a primary internal customer and if they have feedback about formatting, cadence or data captured in the VoC, ensuring CX is receptive and crafting something that speaks to their needs is a good bet to make sure customer voices are being heard.”

In Ronan’s experience, there is a delicate balance of time management for engineers iterating on existing product roadmap and creating new features. His biggest piece of advice to support leaders that are trying to make a case for investment in the product roadmap is to “tie specific areas or problems that we observed in support back to dollar figures. At that point, it is less about asking for product work or new tools to make support’s life easier, but instead, focus the conversation on the potential savings for the business by investing in solving certain issues that the customers were facing.”

Expectation setting with customers helps set support up for success

Simply put, David says this is foundational. “The gap between expectations and delivery is the happiness or frustration zone. In other words, the space between over-delivering and expectations is the happiness zone. The opposite is the frustration zone.”

Getting ahead of this as early as possible, ideally in the sales process, is the best way to ensure the transition process from sales to post-sale is seamless, and there are no surprises. Naturally, sales will want to showcase the best of everything in the business, including support, and anything countering that narrative can make sales nervous. Ronan suggests “it’s best to get out in front of that with clear sales collateral that lays out the expected service levels. Ideally, this material should be created by the service leadership and partner with marketing — it should cover concepts like tiers of service; fee-based upgrades and what benefits they carry; business hours definitions, and information about out of hours or 24/7 support, if it exists in your business.” It should be a resource for salespeople to be able to refer to and answer questions that the potential customer might have.

Support can often reach a threshold during volume spikes in the business, outside of the capacity plan for unanticipated reasons. Eleanor says in normal times, the workforce management planning should account for internal volume spikes. “Learn what those moments are for your business, and don’t let them become a pattern. Make sure, if possible and when necessary, you have a burst capacity staffing plan in place to ensure you maintain SLAs when volumes spike. In these times, also make sure you’re proactively communicating with customers using whatever means your service desk has available. If you know the wait time is well outside the SLA, turn on auto-response emails explaining the situation and make sure customers know how to self-serve as necessary. Proactive communication, with an eye towards ensuring customers know about other resources available to them to self-serve, goes a long way.”

Customer support is a critical factor in boosting retention and reducing churn, but balancing customer needs with company resources and systems is a delicate dance: businesses never want a customer left out to dry, but they also don’t want their agents overloaded with tickets. A great support organization commands the attention of all the stakeholders in the business to ensure that communication internally and externally is aligned and leverages quality customer data and insights to inform all decisions, from staffing to tools and even product roadmap priorities.

— Gina Yocom, Primary Director of Talent

To keep up to date with Primary, don’t forget to sign up for our newsletter.