# Turning customer questions into a content backlog The questions your support team answers every week are a content backlog nobody's mining. Here's how to pull them out and turn the real gaps into pieces worth writing. **Time:** 10 minutes **What you'll have when you're done:** A ranked list of content topics drawn from real customer questions, grouped by theme, with the most-asked and least-covered questions at the top. **Before you start:** - Any AI chat tool you're allowed to use - A batch of real customer questions: support tickets, sales call notes, or a FAQ inbox ---Most content calendars get filled with what the team feels like writing about. Meanwhile the support inbox is full of the same three questions, over and over, and none of them have ever become a piece of content, because nobody's connected the two. This turns a batch of real questions into a backlog ranked by what's actually being asked, not what's easy to write about. --- ## Step 1: gather the raw material, unedited Pull actual customer language, not your team's summary of it. A support agent's ticket notes, sales call transcripts, or a FAQ inbox export all work. Don't clean it up first; the exact phrasing customers use is part of what you're mining, since it tells you the words to use in the content itself. ``` Here are raw customer questions and messages from [source, e.g. "last month's support tickets"]: [paste as much as you reasonably can, 20-50+ if you have them] ``` More is better here than in most prompts in this workflow. A handful of questions produces a handful of themes; a real batch produces frequency, which is the thing you actually need to prioritize by. --- ## Step 2: ask for themes, not a rewritten list ``` Group these into themes: recurring questions or confusions that show up more than once, in different words. For each theme: - How many times does something like it appear? - What's the underlying question, in plain language? - Quote one or two representative examples, in the customer's own words Don't suggest content ideas yet. ``` Holding off on content ideas here matters for the same reason it did in the interrogation-before-spec workflow: jumping straight to "here are 10 blog post ideas" skips past the actual signal, which is which questions are common and which are one-offs. --- ## Step 3: cross-check against what already exists Before turning a theme into a new piece, check whether it's already covered somewhere, and if it is, whether the coverage is actually answering it. ``` Here's a theme from customer questions: [paste it] Here's what we currently publish that's closest to answering it: [paste a link, title, or summary, or say "nothing"] Does the existing content actually answer this question as customers are asking it? If not, what's missing? ``` A theme with existing content that doesn't actually answer it is a rewrite, not a new post, and it's worth knowing the difference before both compete for the same search result or get published as overlapping pieces. --- ## Step 4: rank and turn the gaps into a backlog ``` Given these themes with frequency counts and coverage gaps: [paste the results from steps 2 and 3] Rank them by frequency times gap: most-asked and least-covered first. For the top 5, suggest a specific format (FAQ answer, full guide, short social post) matched to how complex the underlying question actually is. ``` Matching format to complexity is the part worth checking by hand. A question that's genuinely simple doesn't need a long guide, and a question tangled up in edge cases doesn't fit in a three-sentence FAQ answer, no matter how often it's asked. --- ## Worked example Sixty support tickets from a month, pasted in raw. Step 2 surfaced a theme appearing in eleven of them, in different words each time: confusion about why a discount didn't apply to a renewal the way it applied to the original purchase. Step 3 found existing coverage: a pricing page that mentioned renewal terms in one sentence, technically accurate, not actually answering the specific confusion showing up in tickets. Gap: the page never explained the one condition, an upgrade during the term, that caused the mismatch customers were hitting. Final backlog item: not a new page, a specific addition to the existing pricing page addressing that one condition directly, plus a short FAQ entry using the customer's own phrase for the confusion, since that phrase was what several tickets had in common and likely what people search for. --- ## When it breaks **Every ticket looks unique and nothing groups into a theme.** Usually means the batch was too small, or too narrow in date range, to show a real pattern. Widen it, or accept that this particular question set may not have a strong recurring theme yet. **A "theme" turns out to be one person asking the same thing five times.** Frequency counted at the ticket level can overcount a single confused customer. Ask it to note how many distinct customers are behind each theme, not just how many messages. **The suggested format doesn't match how your audience actually searches.** A format recommendation is a reasonable guess, not a research finding. If you have any actual search or site-search data, that overrides the guess every time. **The top-ranked gap turns out to be something you deliberately don't answer publicly.** Pricing exceptions and policy edge cases sometimes fall in this category. That's a real reason to skip a topic, and it's a better reason than "we didn't think of it." --- ## What to try next Run this quarterly, not once. The questions shift as the product changes, and a backlog built from six-month-old tickets will rank yesterday's confusions above today's.

Next