Five Screens for One Task: The Productivity Cost of Fragmented Work

The tolerated pain

A customer phones to ask whether their order can be delivered on Friday. To answer, the person on the phone opens the accounting package to check the order, the stock spreadsheet to see what is on hand, the courier portal to check collection slots, an email thread to find what the customer was promised last time and the team chat to ask the warehouse whether anything has changed. Five screens, three logins and a phone call later, the answer is "probably". The customer has waited, and the next call is already ringing.

Nobody in the business would call this a problem. It is simply how the work is done. Each application was bought for a good reason, each holds part of the answer, and the people who use them have become skilled at stitching the pieces together in their heads. That skill is exactly what hides the cost.

This article is about one task needing many applications. It is related to, but different from, the cost of constant interruptions and context switching, which deals with messages and notifications breaking concentration, and from paying for several applications that do the same job, which deals with duplicate licences. Here the applications may all be necessary. The cost comes from the way the work is spread across them.

Why businesses put up with it

• Each tool solved a real problem when it was bought. The accounting package, the courier portal and the stock sheet were added one at a time, and nobody designed how a single task would move across all of them.
• The switching is invisible. Timesheets record "customer queries" or "order processing", not "time spent finding the right screen".
• Experienced staff are fast at it. Long-serving employees know which tab holds what, so the process looks efficient until they are on leave.
• Integration sounds expensive. Connecting systems feels like a large project, while one more browser tab feels free.

Operational symptoms

• routine tasks that need three or more applications open at the same time;
• staff keeping personal "cheat sheets" of which system holds which field;
• the same customer or order number typed into several search boxes;
• answers given to customers with caveats such as "let me check and come back to you";
• new employees taking months to become productive because they must learn the route between systems, not just the systems;
• errors that appear when a detail is read from one screen and typed into another.

Why switching costs more than the clicks

The obvious cost is mechanical: logging in, waiting for pages to load and searching for the record again. The less obvious cost is mental. Research on task switching has found that people are slower and less accurate immediately after changing task, and that some of this cost remains even when they know the switch is coming. A research editorial summarising studies of multitasking and task switching describes residual switch costs that persist after long preparation, partly because the previous task's settings linger. Each application has its own layout, terms and logic, so moving between five of them to finish one job asks the brain to reconfigure five times.

Fragmentation also shows up in how people experience their day. Microsoft's Work Trend Index report Will AI Fix Work? reported that 62% of respondents struggle with too much time spent searching for information, and 68% said they do not have enough uninterrupted focus time. Those are survey findings about knowledge workers generally, not a measurement of any particular business, but they match what operations managers describe when one task is spread across several tools.

How the pain becomes money

• Labour cost: paid minutes spent navigating, re-searching and re-reading rather than deciding or serving.
• Error and rework cost: details copied by eye between screens are mistyped, and the mistakes surface later as wrong deliveries, wrong invoices or credit notes.
• Revenue and service cost: slower answers mean longer calls, more call-backs and customers who give up and buy elsewhere.
• Scaling cost: because each task takes longer than it should, growth in volume requires more people sooner.
• Onboarding cost: new staff take longer to reach full speed, and the business depends on the few who know every route.

An illustrative cost model

This example is illustrative. The figures are assumptions chosen to show the method, not benchmarks. Picture a Gauteng wholesale distributor with eight people in sales support handling order and delivery queries.

On these assumptions each person loses about 80 minutes a day, or roughly R373 of paid time. Across eight people that is close to R3,000 a day and about R657,000 a year, before counting errors, call-backs or lost orders. If better integration or a single working view removed half of the switching, the recovered time would be worth around R330,000 a year, or roughly two-thirds of one full-time salary on the same assumptions. Your own figures will differ, and the only credible way to know them is to time real tasks.

When it becomes commercially urgent

1. Customer-facing staff cannot answer common questions without putting the customer on hold.
2. Volume is growing and the response is to hire more support staff rather than shorten the task.
3. Errors traced back to copying between screens are showing up in credit notes or write-offs.
4. One or two people are the only ones who can complete certain tasks quickly.
5. A software renewal or new purchase is about to add yet another screen to the route.

What better looks like

Better does not necessarily mean fewer applications. It means the most frequent tasks can be completed from one place, with the other systems feeding that place instead of being visited one by one. That can be an integration that pulls stock and delivery status into the order screen, a shared workspace that surfaces information from several tools, or a small internal app built for one high-volume job. Microsoft, for example, describes its Teams developer platform as a way to bring existing web apps, SharePoint sites and Power Apps into the place where people already work, and to automate processes such as approvals and helpdesk requests. Other platforms offer similar options. The principle matters more than the product: design the route for the task, then decide which tools serve it.

A useful test, described in when copy and paste is really an integration problem, is to look for places where people move information between systems by hand. Those are usually the places where switching is heaviest.

Practical next actions

1. Pick the five most frequent tasks in one team and list every application each one touches, in order.
2. Time a sample of each task and note how much of the time is navigation, logging in and searching rather than judgement.
3. Mark every point where information is read on one screen and typed into another. These are candidates for integration.
4. Decide which system should be the main working view for each task, and which systems should only feed it. The guidance in how to decide what should be the system of record helps here.
5. Fix the highest-volume route first, measure the time saved, and use that evidence before tackling the next one.
6. Check new software purchases against the routes you have mapped, so a new tool shortens a task rather than adding another stop.

Where the mapping shows that staff have become the link between platforms, the pattern in when your staff become the integration between software platforms is likely to be present. Designing integrations around real tasks is the focus of Mellorca's systems implementation and integration work.

Bottom line

When one task needs five applications, the business pays for the journey between them on every repetition. The cost is small per task and large per year, and it grows with volume. Measuring how long common tasks take, and how much of that time is spent moving between screens, usually shows whether the business has a software problem or a design problem, and it is most often the second.

Sources and further reading

• Editorial: Multitasking: Executive Functioning in Dual-Task and Task Switching Situations - PMC
• https://www.microsoft.com/en-us/worklab/work-trend-index/will-ai-fix-work
• Microsoft Teams Developer Platform - Teams