Measuring Digital ROI and Growth

Most ROI conversations about software happen too late, after a system is live and the only available comparison is memory. Your staff will tell you it is better, and they will probably be right, but you will have no way to put a figure on it or to defend the spend to a partner.

Collect four numbers by hand before anything changes, add up the real cost of the build including your own people's time, and split the commitment into stages small enough to cancel.

What counts as a return when software is the spend

A useful return is one you can check yourself. In an SME it almost always takes one of four shapes: hours consumed by a single process, the number of errors somebody has to correct, orders or enquiries lost partway through, and the number of days between sending an invoice and the money arriving.

Leave engagement, awareness and looking more modern out of the calculation, because no decision follows from those numbers. If the figure goes up and you still do not know what to do differently on Monday, it cannot carry the weight of justifying a spend.

Where this measurement sits in the larger sequence of digital work is covered in our digital transformation roadmap for SMEs. Every step on that roadmap gets judged the same way.

Measure the current process before you change it

The baseline has to be collected manually, and one or two weeks is usually enough. Ask everyone involved to record three things: which step they handled, how long it took, and how many times they had to redo it because something was wrong. That comes to a line or two per task over one or two weeks, written in the team's WhatsApp group rather than on a new form, and the person who runs the process is the one asking for it.

The record does not need to be tidy. A spreadsheet tab will do, as long as dates are clear and nothing is reconstructed from memory afterwards. What matters is that the numbers exist before anything changes, because once the new system is in place people compare it against a guess about the old one.

Two weeks of logging usually surfaces something you did not ask for: a step only one person understands, or a duplicate entry that has been running for years without anyone noticing.

The full cost of a build, not just the quote

The figure on the proposal is one line among several. Add the price of each stage, hosting and domain costs, then the hours your own staff will spend reviewing demos, cleaning up old data and learning the new system. Migrating records out of an ageing spreadsheet tends to consume your people's time rather than a developer's.

Add one more line for running it after launch: hosting, monitoring, fixes and small changes. We offer that as an optional monthly service rather than a condition of the build, and the server and domain accounts stay in your company's name, so you can stop paying without losing anything.

Put the total of all those lines next to your baseline and compare them.

The costs that only appear later

Some costs appear in no quote and only make themselves felt after a year. Vendor dependency is the expensive one. If the code, the servers and the accounts are not in your name, every small change has to go through one door, and the price is set at that same door.

Undocumented code does something similar. The next developer inherits an approval step with two stages and no record of why it was built that way, and the second stage gets deleted because nobody wrote down what it was for.

We write the reason for each technical decision next to the code it belongs to, so that two stage approval arrives with its own explanation. The tools underneath are proven and conventional rather than exotic, so whoever maintains the system next is working with things they already know. The code, servers, domain and every account are in your name from day one, so nothing depends on us continuing.

We go through this side of the equation at greater length in the hidden cost of poor software choices.

Four numbers that cover most SME cases

The figures below are placeholders chosen to show the arithmetic. They are not results from any project. Replace them with your own.

Hours returned per week, multiplied by what an hour of that work costs you. Say your admin and counter staff together spend ten hours a week copying orders into a stock spreadsheet. At Rp 50,000 an hour, once you include what the staff actually cost you, that activity is worth Rp 500,000 a week.

Error and rework rate. Count how many orders a month need correcting, how many minutes each correction takes, and add the reshipping costs or apology discounts where they apply.

Revenue a manual bottleneck is holding back. Bookings abandoned because confirmation only arrives the next day, or quotes lost because a competitor replied first. If what you are measuring is reach and inbound enquiries rather than internal process time, the numbers live somewhere else, and we cover that in our piece on why a strong digital presence matters.

Cash cycle time. Work out the average number of days between delivery and payment. Taking five days off that figure changes the cash available to you, and you will feel it even if profit does not move.

Deciding before you spend: when the honest answer is an off-the-shelf tool

Before approving a build, set your custom estimate against the subscription cost of an existing tool over two or three years. Include the switching cost for the day your business outgrows that tool, because the cost is real and it is usually left out.

Custom software is expensive and we will not pretend otherwise. If what you need is invoicing, scheduling or a straightforward online shop, there is very likely a subscription product that does it better and cheaper than anything we would build. When that is the case we will say so and point you at the tool.

Custom work earns its price when the process is genuinely specific to your business, or when the tool you pay for every month has become the reason your staff work by hand.

Staged commitment as a way to limit downside

What you control is how much money depends on the forecast at any one time.

Discovery runs one to two weeks and produces a written plan with a fixed price. If the number does not work once you have read the plan, the plan is still yours, including if you take it to another studio. That makes your first serious ROI estimate cheap to obtain.

Later stages follow the same shape, with scope and price agreed in writing before the stage begins, so stopping after any stage is an ordinary thing to do rather than an argument. Throughout the build there is working software every week and a private preview link from the first week. You check progress yourself instead of reading a status report you have to believe, and if reality is drifting away from the forecast you still have time to stop.

A review rhythm after launch

Re-measure the same numbers at 30, 90 and 180 days, with the same method and one named person who owns them. Without a named owner the measurement quietly stops in month two.

At 30 days your staff are still learning, so read that measurement as a map of where people are working around the new system rather than a verdict on it. Those are usually process fixes. By 90 days the figure can be trusted. The last reading, at 180 days, tells you whether the improvement held or whether people have drifted back to the old spreadsheet.

If the numbers disappoint, there are three sensible responses. Fix the process, since the software is sometimes right and the order of work around it has not caught up. Change the software, if a step has been forced into the wrong shape. Or stop paying for the monthly operation, since the accounts are in your name and nothing obliges you to continue.

Work the numbers through with us

Pick the single process that consumes the most staff time, describe it to us on a free first call, and we will work the numbers through together.

If the right answer turns out to be a subscription tool, we will show you which one. If it is not, Discovery gives you a written plan and a fixed price, and both stay yours whether or not you continue with us.

Message us on WhatsApp at +62 851-1769-7889 or email hello@arktik.id.