If you run a small business and have decided you should be better with your numbers, the temptation is to buy a dashboard tool and track everything. Do not. Start with four weeks of deliberately narrow work, because a small number of well-understood measures beats a wall of charts nobody reads.
Week 1: Build one honest weekly number sheet
One spreadsheet, one row per week, and no more than eight columns. Suggested starting set:
- Revenue
- Number of orders or jobs
- New customers
- Enquiries or leads
- Cash in the bank at week end
- Marketing spend
- One operational measure that matters to you (delivery days, occupancy, utilisation)
Fill it in for the last 26 weeks from your existing records. This backfill is the most valuable few hours you will spend, because a single week of data tells you nothing and 26 weeks tells you what normal looks like.
Week 2: Learn your normal range
For each column, work out the average and, more importantly, the highest and lowest typical weeks. You are looking for the band your business normally sits in.
This is the foundation of everything. Once you know revenue normally sits between £8,400 and £11,900 a week, a £9,100 week stops being a worry and a £13,200 week becomes genuinely interesting. Most small business stress comes from reacting to ordinary variation as though it were signal.
Week 3: Calculate three ratios
- Cost per new customer = marketing spend ÷ new customers, per channel if you can split it
- Conversion rate = orders ÷ enquiries
- Repeat rate = customers who bought again within 12 months ÷ customers who bought
These three tell you where your money is being wasted, where your process leaks and whether you are building anything durable. Almost every owner finds at least one surprise here.
Week 4: Look at the spread, then run one test
Take your order values and sort them. Look at the smallest quarter and the largest quarter. Are the small orders profitable once you count your time? This one exercise has changed the pricing of a very large number of small businesses.
Then pick the single cheapest test you can run - a price change on one product, a different headline on your ads, a follow-up email to lapsed customers - and run it for four weeks with the numbers recorded. You now have a habit, not a project.
What to deliberately ignore for now
Vanity metrics (followers, impressions, page views without conversion), any dashboard tool, statistical software, and anything requiring more than a spreadsheet. All of it can come later; none of it helps yet.
The mindset that makes the difference
Two questions, asked every time a number moves: is this outside my normal range? and what would I do differently if it stayed this way? If the answers are "no" and "nothing", carry on with your week. That discipline is worth more than any software.
Going further
When you are ready to add proper testing, forecasting and confidence intervals, the Statistics for Business course at London School of Business UK is practical and built for people running things rather than studying them. Get in touch.