You track sales weekly and cash monthly, but if I asked what your technology spend achieved last quarter, could you answer with a number? For most owners the honest answer is no, and it is not their fault. Nobody ever handed them a short list of technology KPIs for business use, so tech remains the one budget line judged entirely on feelings: it "seems fine" until the day it visibly is not.
The result is predictable. Tech spending gets approved when a vendor tells a good story and cut when cash is tight, with no feedback loop in between. Systems that nobody uses keep getting paid for. Systems that quietly save hours get no credit. And the IT conversation stays stuck at "is it broken?" instead of "is it earning?"
You do not need a dashboard project or a data team to fix this. You need four numbers on one page, reviewed monthly, each with a threshold that tells you when to act. Here is the scorecard I recommend, and how to get each number cheaply.
Why most tech metrics are useless to an owner
First, what to ignore. Vendors and IT teams will happily report server CPU utilization, ticket counts, story points delivered, or page views. These are operating metrics for the people running the systems. They are not decision metrics for the person paying for them.
An owner-grade KPI has to pass three tests:
- It connects to money or hours. Directly, not through three assumptions.
- You can get it monthly without a project. If measuring it takes a consultant, it will not survive three months.
- It has a red-flag threshold. A number without a "so what" level is trivia.
Four metrics pass. Here they are.
KPI 1: Uptime of your critical systems
The question it answers: can my business operate?
Pick your two or three genuinely critical systems, the ones where an outage stops sales or operations: POS, order system, website checkout, WhatsApp line. Track the percentage of business hours they were actually usable this month.
How to measure cheaply: free or near-free uptime monitors (UptimeRobot and similar tools) ping a website or server every few minutes and email you a monthly percentage. For offline systems like a POS, keep a simple downtime log: date, duration, cause. Whoever experiences the outage records it in a shared note. Crude, but honest.
Red flag: below 99 percent during business hours. That sounds strict, but 99 percent of a 12-hour retail day still means roughly 3.5 hours of outage a month. If you are below it, the fix conversation is usually specific: hosting, a flaky integration, or a vendor's release discipline. And whatever the number, you should already know what happens during the downtime you cannot prevent; that plan is the subject of business continuity for SMEs.
KPI 2: Adoption of the systems you pay for
The question it answers: is anyone actually using what I bought?
This is the most embarrassing and most valuable number on the scorecard. For each system with a subscription or maintenance cost, track active users this month divided by intended users. A CRM bought for 12 salespeople that 3 actually use has 25 percent adoption, which means 75 percent of the subscription is burned.
How to measure cheaply: most SaaS admin panels show last-login dates; that is a two-minute monthly check. For custom systems, ask your developer for a monthly active-user count; it is a trivial query. For shared tools, ask team leads to name who used it this week and cross-check occasionally.
Red flag: below 60 percent after the first three months. Below that line you have one of three problems: the tool is wrong, the training never happened, or the process does not require the tool (people can still do the job the old way, so they do). Each has a different fix, but all are cheaper than continuing to pay for shelfware. In my experience, low adoption is the single most common way SMEs waste tech budget, more than overpriced vendors and failed projects combined.
KPI 3: Technology cost per order (or per transaction)
The question it answers: is tech spend scaling sensibly with the business?
Add up all monthly technology costs: subscriptions, hosting, maintenance retainers, payment gateway fees, a fair slice of any IT salary. Divide by the number of orders or transactions that month. A seller doing 4,000 orders on Rp 12 million of monthly tech spend has a tech cost of Rp 3,000 per order.
The absolute number matters less than its trend. Healthy pattern: the ratio falls or holds as volume grows, because good systems have mostly fixed costs. Unhealthy pattern: the ratio climbs while volume is flat, meaning tools are accumulating without anything being retired.
How to measure cheaply: one spreadsheet row per month. Your bookkeeper already has every input.
Red flag: the ratio rising for three consecutive months without a deliberate investment you can name. A spike you chose, a new system going live, is fine. Drift is not. The usual culprits are subscription creep and oversized cloud bills, and both respond well to a periodic cleanup of the kind I described in cutting cloud infrastructure costs without breaking things.
KPI 4: Hours saved by automation this month
The question it answers: is technology giving my team time back?
Every automation you deploy, automatic report generation, order sync, payment matching, template-based drafting, replaces a task someone used to do by hand. Estimate the hours once, honestly, when the automation goes live ("this used to take Sari 6 hours a week"), then count it monthly while the automation keeps running.
Yes, it is an estimate. It is still transformative, because it makes invisible savings visible. A back-office automation saving 40 hours a month is worth roughly a quarter of a full-time salary; without this number, that value evaporates from every budget conversation, and the next automation proposal gets judged as pure cost.
How to measure cheaply: a running list: automation, hours saved per month, date verified. Re-verify the estimates twice a year by asking the team.
Red flag: the number has not grown in six months. Not because every month needs a new robot, but because a flat line usually means nobody is looking for the next repetitive task, and in any real SME back office there always is one. Reconciliation alone is often worth days, as I showed in payment reconciliation automation.
The one-page scorecard in practice
Put the four numbers in a table, one row per month:
| Month | Uptime (critical systems) | Adoption | Tech cost / order | Hours saved |
|---|---|---|---|---|
| Jul | 99.4% | 71% | Rp 3,100 | 62 |
| Aug | 98.2% | 68% | Rp 3,350 | 62 |
Review it in thirty minutes a month, ideally in the same meeting where you review sales. Two rules make it work. First, every red flag gets one named action, not a discussion, an action with an owner. Second, resist adding metrics. The moment this becomes a ten-row report, it stops being read, and the entire value of technology KPIs for business owners is that they fit on one page you actually look at.
The takeaway
Technology stops being a leap of faith the month you start scoring it: uptime tells you whether the business can operate, adoption tells you whether purchases are real or shelfware, cost per order tells you whether spend scales sanely, and hours saved tells you whether automation is earning. Build the scorecard this week, backfill last month from records you already have, and put a recurring thirty-minute review on the calendar. If you would rather set it up with someone who has built these scorecards inside real companies, that is a conversation I am glad to have.