Skip to content
feelingstream
//Article

19 January 2021/Terje Ennomäe

Cut avoidable inbound calls with conversation analysis

Avoidable inbound calls reduced with conversation analysis

Every contact centre carries a share of calls that never needed to happen. Customers ring to track a parcel they could track online, or to query a charge that a clearer bill would have explained. Each of those calls costs money and time — and, with the right changes, could have been avoided entirely.

The catch is finding them. Avoidable calls are scattered across thousands of contacts and hidden inside broad category tags. Going through calls one by one to identify which are avoidable, and why, is impossibly slow. So most companies know their volumes are high but cannot say how much of that volume is avoidable, or what to fix.

Conversation analysis changes that. By analysing every call and grouping them by topic, you can quantify avoidable contacts and trace them to a root cause worth fixing.

Avoidable calls add up

Each call to customer service is a cost, so each avoidable one is money and time the company need not have spent. As a rough guide, when 10–25% of contacts turn out to be avoidable, it is time to act.

That figure is rarely one big problem. It is many small ones — a confusing invoice line here, a hard-to-find self-service feature there — and they add up. Two common patterns:

  • Logistics: parcel tracking. Tracking is usually available online, yet customers still call. People will not stop calling just because you want them to; the fix is a more user-friendly tracking system, or a proactive message with the relevant information before they reach an agent.
  • Telecom: bill charges. A large share of calls are about charges on monthly bills. When analysis shows customers keep asking about the same things, the service descriptions, self-service and bills themselves need to be made clearer.

Where the issues hide

Today most companies only know what customers called about from agent notes and predefined categories. Those categories are usually too coarse for real analysis, and agents cannot fix the underlying problem anyway — because it lives in the process, the self-service flow, or an uninformative page, not in the call itself.

The only reliable way to make an impact is to analyse the avoidable calls in depth. Doing that by hand is unrealistic at scale, which is why it so rarely happens.

Automated analysis that goes deep

Feelingstream's platform transcribes every call and detects topics automatically, in far more detail than manual agent tagging allows. It looks for patterns and groups calls into a topic map, so quality managers, product owners and service owners can see together what customers really call about — and then decide which of those are avoidable.

Frequent topics branch into sub-topics: "bills" splits into specific charges, offers and services. Going in depth on those sub-topics is what reveals the precise change a customer needs.

The arithmetic is compelling. Suppose the topic map shows 20% of calls are about billing, and analysis finds more than half of those are avoidable. That is 10% of total volume you could work to eliminate. Each individual fix may look small, but the small percentages accumulate into a large reduction.

The benefits go beyond cost

Removing 10–25% of avoidable calls does more than lower cost. It changes how the whole operation feels and performs:

  • Agents spend less time on trivial, repetitive questions.
  • Shorter queues mean wait times fall and satisfaction rises.
  • Self-service and processes, improved around real customer needs, become more usable — so fewer frustrated customers call in the first place.
  • Simpler processes lift employee satisfaction too.
  • Agents get room to work on sales skills, relationships and retention.

This is not one-off work. Markets, competitors and your customer base all change, and new issues appear as you fix old ones. But with continuous topic analysis you keep finding where the business can run more efficiently. See our guide to efficiency with AI for the wider approach.

Frequently asked questions

What counts as an avoidable inbound call?

A call the customer would not have needed to make if a process, bill or self-service flow were clearer — for example calling to track a parcel that is already trackable online.

How do you find avoidable calls at scale?

By transcribing every call and detecting topics automatically, then grouping calls into a topic map. That lets you quantify how much volume each avoidable issue represents instead of sampling manually.

Why aren't agent notes enough?

Predefined categories are usually too coarse for real analysis, and agents cannot fix the root cause — it typically lives in a process, page or self-service flow rather than in the call itself.

What impact can reducing avoidable calls have?

Beyond direct cost savings, it shortens queues, cuts wait times, improves both customer and employee satisfaction, and frees agents to focus on sales and retention.

Where to go next

Curious how many of your inbound calls are avoidable? We will find them in your own conversations. Book a demo.