Monday, 8.30am. The coffee is still too hot to drink and the inbox says 37 unread. Some are form submissions that say only "please call me". Two are the same person, who wrote again on Sunday because nobody replied on Saturday. One is a complaint. One is a large job that will be booked with somebody else if you do not call before ten.
Most small businesses deal with this by starting at the top and working down. It feels productive and it is the worst order possible, because the easy messages get fast replies while the important ones wait behind them.
A Monday enquiry backlog is not a sign that something is broken. It is a predictable weekly event, which means it can have a predictable routine.
The routine, step by step
This takes a disciplined first hour. Adjust the times to your volume.
- 8.30 to 8.40: read, do not reply. Scan every message and give each one a label: urgent, new work, existing customer, question, junk. Delete or archive the junk immediately.
- 8.40 to 8.50: merge duplicates. Find people who wrote twice or used both the form and chat. Keep one thread per person, with all their messages together.
- 8.50 to 9.10: urgent and existing customers. Problems with current jobs, complaints and anything time-sensitive. Call where possible; a phone call on Monday morning repairs a lot.
- 9.10 to 9.40: new work, highest value first. These are the enquiries competitors are also answering. Call or send a proper reply, not an acknowledgement.
- 9.40 to 10.00: questions, in batches. Group similar questions and reply with an edited template for each group.
- 10.00: feed the gaps. Note any question asked more than twice and add the answer to your website or assistant today.
The two rules that matter most are the first and the last. Reading everything before replying stops you giving the best hour of your week to the least important messages. Feeding the gaps means next Monday's pile is smaller.
Why "please call me" is the slowest message in the inbox
The real time sink is not the number of messages. It is how much each one tells you. Compare the same enquiry arriving two ways.
| Bare form submission | Captured with details | |
|---|---|---|
| What arrives | Name, phone number, "please call about a quote" | Name, phone, email, job type, postcode, rough size, preferred call time, the questions they already asked |
| First step on Monday | Call to find out what they want | Decide whether you do this job in that area |
| If they do not pick up | Voicemail, try again later, still know nothing | Send a specific email with a price range or next step |
| Can you prioritise it? | No, every form looks the same | Yes, by job type and value |
| Typical back and forth before a quote | Several contacts | Often one |
The captured version is what a website assistant produces when it asks a few follow-up questions before taking contact details. Nothing magic happens. It simply asks what you would have asked, at the moment the customer is keen to answer. For more on getting that data into the tools you already use, see how chatbots feed your CRM with clean lead data.
A worked example you can redo with your own numbers
Take a hypothetical Monday with 30 weekend enquiries. Replace the figures with your own after timing yourself.
Without captured details. Say each enquiry needs, on average, 3 minutes to read and understand, plus 7 minutes to call, leave a message or ask clarifying questions. That is 10 minutes each, or 300 minutes. Five hours, and most of those people still do not have an answer by lunchtime.
With an assistant answering and capturing over the weekend. Suppose 10 of the 30 were simple questions (opening hours, areas covered, whether you do a particular service) that were answered at the time and need no reply. The remaining 20 arrive with details attached. Reading drops to 1 minute, because the job is described, and replying drops to 4 minutes, because you can go straight to a price range or a visit. That is 5 minutes each, or 100 minutes.
Five hours becomes a bit under two, in this example. Your ratios will be different, which is exactly why timing two real Mondays is worth doing. The chatbot ROI calculator can help you put a value on the hours.
Templates that do not sound like templates
Batched replies save time, but a reply that obviously ignores what the customer wrote does damage. Keep a handful of starting points and always edit the first line.
- Price question: open by restating their job in their words, give the range or what affects the price, then one clear next step.
- Quote request with details: confirm what you understood, propose a specific time for a call or visit, ask one missing question at most.
- Existing customer update: lead with the answer to their actual question, then any logistics.
- Enquiry you cannot help with: say so in the first sentence, and if you know someone who can, mention it.
Our guide to what to do with the first lead your chatbot captures goes further on replies that keep momentum.
Reducing next Monday's pile
A backlog is partly a content problem. Every question that appears on Monday and could have been answered on Saturday is one you are paying for twice.
Keep a running note during step six. After four Mondays you will have a clear list of the questions that keep coming back. Put those answers on the website and into the assistant's knowledge, and check the unanswered questions list it records over the following weekend. Where enquiries need to reach a particular person quickly, a webhook can post them into a CRM or a Slack channel as they arrive; the integrations page lists the common routes.
It also helps to understand why weekend messages build up in the first place, which is covered in the weekend enquiry problem.
Next Monday, start here
Print or save the six-step routine and follow it once exactly, even if it feels slow. Time how long the backlog takes and count how many messages were bare "please call me" requests. That number is the size of the problem that better capture solves.
Then pick the three questions that appeared most often and write proper answers for them before Friday afternoon.