Use case · Choosing what to improve
Decide which website or marketing improvement to work on next
You don't need another list of "20 ways to improve your website". You need to know which one thing to do next. The answer is usually already in your inbox.
By The Dance teamPublished 5 min read
- Time it takes
- One 60–90 minute session, then 15 minutes a month
- You'll need
- Recent customer emails, texts, DMs and call notes
- Google Search Console access (useful, not required)
- Any list of website ideas you already have
- What you'll have at the end
- A ranked shortlist, one improvement chosen with a reason, and a number to check in four weeks.
The short answer: choose your next improvement from evidence, not from a best-practices list. Collect what customers ask you, what people searched to find you, and what you notice in enquiries. Turn each pattern into a possible improvement and score it on three things: how many customers it affects, how close it is to a sale, and how easy it is. Then do the top one, and only the top one, before choosing again.
Generic advice ("add testimonials", "improve your speed", "start a blog") isn't wrong. It's just not ranked for your business. Your customers have already told you what's missing. This workflow writes it down.
Where the evidence already lives
You likely have four sources. You don't need all of them to start.
1. Customer questions
The questions people ask before they buy are the clearest signal of what your website doesn't explain. Look through the last month of emails, texts, social messages and call notes. Write each question down in the customer's own words.
Look for: the same question asked three or more times, and questions asked right before someone booked or decided not to.
2. Searches people used to find you
If your site is connected to Google Search Console, its Performance report shows the searches (queries) your site appeared for, how often it was shown (impressions) and how often people clicked. Export the top queries for the last three months.
Look for: searches that mention a service or place you serve but get few clicks, and searches for things you offer but don't have a page about.
3. Patterns in enquiries
What do people ask for when they contact you? Which services come up most, and which ones never come up even though you'd like more of them?
Look for: a service you want more of that nobody asks about. That's often a page that doesn't exist or doesn't say enough.
4. What others in your area answer
Look at two or three competitors' websites and your own Google Business Profile. Don't copy them. Note what a customer can learn there that they can't learn on your site.
Turn evidence into possible improvements
Each pattern suggests a specific change. Write them as changes, not as topics.
| Evidence | Possible improvement |
|---|---|
| "Do you cover Oak Park?" asked six times this month | Add a service-area section listing every town you cover |
| Search impressions for "water heater replacement" but few clicks | Create a dedicated water heater replacement page |
| People ask "how much?" before almost every booking | Explain how pricing works on each service page |
| Nobody asks about your commercial cleaning, which you want to grow | Give commercial cleaning its own page with who it's for |
| Competitors show real photos of finished jobs; you show stock images | Replace stock photos with your own on the three main pages |
Here are the last 20 questions our customers asked, copied from our emails and texts: [paste them]. Group them by topic, count how many fall in each group, and for each group tell me whether our website answers it clearly. Here is the text of our main pages: [paste]. Don't write answers. Just show me the gaps.
AI is genuinely useful here. It can sort a messy pile of messages into groups in seconds. It only sees what you paste in, though, and it can't tell which questions came from people who went on to buy. You know that.
Score the ideas
Score each idea from 1 to 3 on three questions, and add them up.
| Improvement | Customers it affects (1–3) | How close to a sale (1–3) | How easy (1–3, 3 = easiest) | Total |
|---|---|---|---|---|
- Customers it affects: 3 if most visitors would notice, 1 if only a few would.
- How close to a sale: 3 if it's about booking, pricing or availability. 1 if it's general information.
- How easy: 3 if you could do it this week, 1 if it needs new photos, other people, or money.
The highest total wins. If two ideas tie, pick the one closer to a sale.
| Improvement | Affects | Close to sale | Easy | Total |
|---|---|---|---|---|
| Service-area section (Oak Park, Evanston, Skokie) | 3 | 3 | 3 | 9 |
| Water heater replacement page | 2 | 3 | 2 | 7 |
| Explain call-out fee on each service | 3 | 3 | 2 | 8 |
| Replace stock photos | 2 | 1 | 1 | 4 |
| Start a monthly blog | 1 | 1 | 1 | 3 |
The service-area section wins. It answers the most-asked question, it's right at the point of contact, and it's a 20-minute change. The call-out fee explanation comes next month.
Do one thing, then check
Commit to the top item, and write one sentence about what you expect to change. For example: "Fewer calls asking whether we cover their town, and more enquiries from Evanston."
Note the date you published it. Four weeks later, look at the same evidence again. Did that question stop coming up? Did enquiries for that service or place change? Then score the list again, because new evidence will have arrived.
What "next" usually shouldn't be
- A full redesign. If customers can't find your prices, a new look won't fix it. A redesign is sometimes right, but rarely as the next step.
- Chasing searches you can't serve. Ranking for a service you don't offer, or a town you don't cover, brings visitors you'll have to turn away.
- Content for its own sake. A blog post nobody asked for is less useful than one clear answer to a question ten customers asked.
Limits worth knowing
- Small numbers bounce around. If you get 20 enquiries a month, a change from 20 to 24 might be chance. Look for a pattern over two or three months.
- Search Console shows how you appear in Google, not why someone did or didn't call.
- No tool, AI or otherwise, can promise that an improvement will bring customers. The goal is to make better bets, more often, and check them.
Keep going
- ArticleHow to prioritize website improvements when you have one hour a weekOne hour a week is enough to keep a small-business website improving, if you spend it on the right thing. Here's a routine, a scoring worksheet, and a worked example.
- Use caseTurn customer demand into useful service pagesThe best service pages answer what someone wants to know right before they contact you. Here's how to find the service worth a page, and build it from your own facts.
- ArticleHow to measure whether AI is helping your small businessYou don't need an ROI model. You need a dated baseline, four honest numbers, and a clear sense of what you can and can't put down to AI.