Answer
What should a small business automate first?
The failure that leaves no evidence. Everything else is already visible enough to fix without new software.
The failure nobody can see. Start with unanswered enquiries — the ones that arrive out of hours or while everyone is busy — because they leave no record and so are never prioritised. Visible problems get fixed without software.
Sequencing gets decided by volume of complaint. Whatever the team grumbles about most becomes the first automation project, and because complaint volume tracks visibility rather than cost, the first project is usually aimed at something that was already being managed. It produces a modest improvement to a problem that had a person watching it, and it consumes the appetite for change that a bigger fix would have needed.
The better ordering rule is to look for failures that leave no evidence. A missed call at eight in the evening, a web form that went to a mailbox nobody monitors, a message on a channel the business set up two years ago and forgot, an enquiry that arrived while both people were on jobs. Each of these is a complete loss and none of them appears anywhere. No report is short, no customer complains, and the business's own sense of how it is performing is assembled entirely from the interactions that worked.
That silence is why these persist. Every visible problem is subject to ordinary management pressure and gets improved eventually, even clumsily. An invisible problem is not competing for attention, because it never enters the competition. So the highest-return first project is almost always the one that makes a silent failure audible, and quite often the making-audible is worth more than the automating.
Which suggests a first project that costs less than most businesses expect: instrument before you automate. Route every channel into one place, timestamp arrival, and record what happened to each enquiry, including nothing. A week of that data usually reorders the priority list entirely, and it is the same data that later proves whether the automation worked. Businesses that skip it end up with an automation whose benefit nobody can demonstrate and which is therefore vulnerable to the first person who dislikes it.
Once the silent failure is visible, the automation that addresses it is generally simple and cheap: an immediate acknowledgement on every channel at every hour, with a named owner and a deadline behind it. It is unimpressive to demonstrate and it is the single change that most reliably converts into revenue, because the enquiries it saves were being lost completely rather than handled slowly.
The sequencing after that follows the same logic rather than a technology roadmap. Next comes wherever work is stationary for the longest without anyone noticing — the quote that was ready on Tuesday and sent on Friday, the customer waiting on a detail nobody chased. Those are also handoffs, they are also cheap to address, and they are also invisible until something counts them.
Businesses automate the thing people complain about, which by definition is the thing that was already getting attention.
Answer Production Engine, Context Theory
Related questions
Should we fix the process before automating it?
Fix it enough to describe it, then automate. Waiting for a clean process is how automation projects stall indefinitely, because the process is never clean and the mess is often exactly what the automation handles. What you cannot skip is being able to say what should happen when an enquiry arrives — who owns it, by when, and what happens if they do not. Automating an undecided process just makes the indecision faster.
Is it worth automating if we only get a few enquiries a week?
At low volume the case is about loss rate rather than time saved, and it is often stronger. A business receiving a handful of enquiries a week cannot afford to lose any of them, and the loss rate on unmonitored channels is not lower because volume is low. The time-saving argument needs scale; the not-losing-anything argument does not.
METHOD
Every figure below carries its source and the date it was verified. Nothing on this page is asserted.
The numbers on this page.
| What | Value | Specific to |
|---|---|---|
| Firms that never responded to a web enquiry at all | 23% | Category-wide |
| Average B2B first-response time | 42 hrs | Category-wide |
| Leads cold past 5 minutes | 93% | Category-wide |
Oldroyd, McElheran & Elkington, "The Short Life of Online Sales Leads", Harvard Business Review (March 2011) · hours · 1.25M inbound leads across 2,241 US firms · verified
2026 speed-to-lead benchmark · derived: 100% − 7% responding within five minutes · verified
What is specific to this page.
| Kind | Claim | Check it against |
|---|---|---|
| Workflow | Automation sequencing driven by complaint volume targets whichever failure is most visible, and visibility tracks who is watching rather than what the failure costs, so the first project usually improves something already being managed. | The business's own list of proposed automation projects, ranked by who raised each and against what evidence. |
| Software | A business's sense of its own performance is assembled entirely from interactions that completed, because unanswered calls, unmonitored form destinations and abandoned channels generate no record in any system it reviews. | An inventory of every published contact channel, checked for where each currently delivers and whether anyone monitors it. |
| Workflow | Instrumenting arrival — routing every channel to one place, timestamping receipt and recording the outcome including none — usually reorders the priority list within a week and supplies the baseline that later proves whether an automation worked. | One week of arrival-and-outcome logging across all channels, compared with the priority list held before it. |
| Procurement | At low enquiry volume the case for automation rests on loss rate rather than time saved, and loss rate on unmonitored channels does not fall with volume, so the argument that scale is required applies only to the time-saving half. | The share of enquiries receiving no response, computed separately for the business's lowest-volume channels. |
Each row would be wrong on another industry's page. Where a sourced figure exists it is in the table above instead; these are the constraints that shape the work and do not happen to be numbers.
Start with the measurement.
Reading about a benchmark is not the same as knowing your own number. The audit produces yours, measured rather than estimated.
$497 · delivered in 5 business days · credited against month one