How to analyse customer feedback from every channel

Know your channels

Learning how to analyze customer feedback starts with knowing where it comes from, because each channel tells you something different and is biased in its own way.

• Public reviews (Google Maps, app stores): unsolicited, public, written for other customers. Strong on what people care enough to mention; skewed towards memorable experiences.
• Complaints and support contacts: unsolicited and problem-focused. Strong on what went wrong; silent on what went well.
• Surveys (CSAT, NPS, post-visit): solicited and structured. Strong for tracking a score; comments are shaped by the questions asked.
• Social media: unsolicited, public, short. Good for spotting sudden issues; hard to tie to a location.
• Staff feedback: what customers say in person, filtered through the person who heard it.

Use one taxonomy for all of them

Classify every channel against the same set of themes and sub-themes. That is what lets you see that billing complaints are rising in reviews and in support contacts at the same time. Report each channel separately, though: combining them into one number lets a change in channel mix look like a change in experience.

The analysis method

1. Collect each channel with its date, location (if any), score (if any) and full text.
2. Split each comment into mentions of separate topics.
3. Classify every mention into the shared taxonomy, with its sentiment.
4. For each topic and channel, measure volume, negative share and change since the last period.
5. Break the top issues down by location and month.
6. Read examples behind every finding.
7. Give each rising issue an owner, a next step and a date to check.

Survey comments need care

Open-text answers to a survey question are shaped by the question. Ask "What could we improve?" and you will get complaints even from happy customers. Analyse survey comments alongside their scores: what do detractors mention that promoters do not? That comparison is more informative than the comments alone.

Look for agreement across channels

An issue that shows up in reviews, support contacts and survey comments is almost certainly real. One that shows up in a single channel may be real, or may reflect that channel's bias. For big decisions, check at least two channels before acting; for small, cheap fixes, one strong signal is enough.

Report so people act

• Lead with the finding, not the method: "Billing complaints rose from 9% to 15% of reviews at mall stores since July."
• Show the evidence: two or three comments in the customer's words.
• Name an owner and a next step.
• Keep the same format every month, so changes are obvious.

Pitfalls

• Treating all channels as one. Mix changes look like experience changes.
• Counting reviews instead of mentions, and losing half of every mixed review.
• Changing categories every quarter, which breaks every trend.
• Reporting numbers without the words behind them.

An example across channels

A retailer sees complaints about returns rising in Google reviews. Before changing the returns policy, the team checks other channels. Support contacts tagged as returns rose over the same weeks; survey comments from detractors mention refunds taking too long more than they did a quarter earlier. Three channels agree, and they agree on the specific issue: not the policy itself, but the time refunds take to arrive. The fix goes to the finance process owner, not the store teams.

Without the cross-check, the obvious fix would have been to retrain store staff on the returns policy, which would have changed nothing customers were complaining about.

How often to analyse

• Weekly: a quick look for sudden spikes, especially after launches, policy changes or incidents.
• Monthly: the full view by topic, location and channel, with owners and actions.
• Quarterly: review the categories, the measures and whether actions worked.

Measures that travel across channels

Scores do not compare across channels: a 4-star review, a CSAT of 4 and an NPS of 8 measure different things. Topic measures do. The share of feedback in each channel that complains about billing, and how that share changes month to month, can be compared even when the scales cannot. Report scores for each channel on its own, and use topic shares to compare channels.

Doing this with Falcon

Falcon does this for your Google Maps and Google Play reviews. Every review is classified into themes, sub-themes and issues from your own taxonomy, with the sentiment of each mention; specific issues are found by clustering what customers wrote, each with a short summary; and every count opens the reviews behind it, with the supporting quote highlighted.