A spreadsheet-and-notebook setup that is "good enough" usually costs an R&D team hours of scientist time every week, spent re-typing results from one place into another and assembling reports by hand. Those hours rarely show up in a budget line, which is why the setup looks free. At Rogers Corporation, scientists reported spending 40% of their time searching for past experiments and compiling reports. At Cabot, entering the data for a single service request in one elastomers lab took more than six hours.
This piece is for the R&D leader or lab manager who has heard "what we have works" and wants to show what it costs. It covers where the hours go, what transcription does to data quality, and a way to count the cost for your own team.
A setup that works because people hold it together
When a team says its data setup is good enough, it usually means results eventually reach the people who need them. That is a statement about the outcome. It says nothing about who carried the data there.
In most spreadsheet-based labs, scientists are the connection between systems. A result is read off an instrument, typed into a notebook, copied into a team spreadsheet, and pasted into a slide for the project review. Each step works. Each step also takes a person's time, and the setup only keeps working while those people keep doing it.
That is the useful distinction for a business case: a setup that functions is different from a setup that runs without manual effort. Finance hears the first. The cost sits in the second.
Where the hours go: re-entry, reconciliation and report assembly
The time disappears into three kinds of work, and the published customer accounts on uncountable.com show each of them.
Re-entering results is the most visible of the three. Cabot's elastomers team was doing manual data entry into a homegrown database, work the case study describes as "time-consuming transcription." Michelle Shea, Lab Manager at Cabot, put the before figure at more than six hours of her team's time to enter the data for a single service request. After moving to Uncountable, the same request takes three hours on average.
Reconciling data from different sources takes the next share. Before Uncountable, SCG Chemicals' Digital Manufacturing team relied on Excel, Minitab and Python, and gathering data from scattered sources was a major pain point. SCG now reports 50% less time spent mapping data from various sources. Ripple Foods reports a 30% reduction in time spent on data reconciliation.
Report assembly is where the hours are easiest to underestimate. At Rogers Corporation, data was spread across Excel spreadsheets, PowerPoint presentations, handwritten lab notebooks and email chains, and scientists "frequently spent days, if not weeks, compiling reports." After moving to Uncountable, Rogers scientists saved 40% of the time they had spent searching for past experiments and compiling reports.
None of this work produces a new result. It moves existing results from one format to another.
What does re-typing do to data quality?
Re-typing introduces errors, and checking by eye catches fewer of them than most teams assume. In a controlled study of data entry methods, Barchard and Pace (2011) found that visual checking was not significantly better than entering the data once with no check at all. Visual checking produced 2,958% more errors than double entry, roughly 30 times as many.
The study used research data entry, not a formulation lab, so the exact rates won't transfer. The direction does. A value typed twice has two chances to go wrong, and a lab that relies on someone reading back a spreadsheet is relying on the method that study found least reliable. The cost then shows up later: a repeated experiment, a model trained on a wrong value, or a decision made from a report nobody can trace back to the raw result.
Count it for your own team
The cost of doing nothing is a number you can estimate in a week. Ask each scientist to log, for five working days, the time spent on three tasks only: re-typing results that already exist somewhere else, reconciling versions of the same data, and assembling reports or slides by hand.
Then run the arithmetic. As an example, take a team of 30 scientists who each log three hours a week on those tasks:
- 30 scientists x 3 hours = 90 hours a week
- 90 hours x 46 working weeks = 4,140 hours a year
- At about 1,840 working hours per scientist per year, that is the time of roughly two and a quarter full-time scientists
Multiply the annual hours by your own loaded hourly cost to get a figure finance will recognize. Three hours is a conservative input. Ripple Foods' VP of R&D, Andy Seaberg, estimates Uncountable has saved "at least a half-day of work per week per scientist," which suggests the manual load in many labs is higher.
Four checks before you call it good enough
These tests take an afternoon and give a champion something concrete to bring into a budget conversation.
- Pick a result from 18 months ago and time how long it takes to find the formulation, conditions and raw instrument file behind it.
- Ask who built the last project review deck, and how many hours it took. Then ask how many of the numbers in it were typed rather than pulled.
- Trace one value from the instrument to the final report and count how many times a person entered it.
- Ask what happens to the team spreadsheet when the person who maintains it is on leave.
If the answers are "a few minutes," "almost none," "once" and "nothing changes," the setup may well be good enough. Most labs get different answers.
What this means for R&D leaders building a business case
- Which of your scientists' weekly hours go to moving data rather than producing it, and has anyone logged them?
- When finance asks what happens if you wait a year, can you answer in hours and in repeated experiments?
- Which results would you trust less if you knew how many times they had been typed?
- Who on your team is the setup quietly depending on?
Entering a result once and reporting from it
Uncountable is used by 175+ companies in chemicals, advanced materials, food and beverage, and pharma and biotech. Its ELN captures experiments in structured, queryable form with graphing, reporting and statistics built in, and its LIMS for R&D logs test results against the samples and experiments that produced them, flagged in or out of spec on entry. Cabot runs 75+ instrument connections on the platform and reports a more than 50% reduction in time from testing completed to results delivered. For more customer accounts, see the Uncountable customers page and case studies.
.png)
.png)
.png)
