The Problem
There are no commercially available and appropriately priced winery laboratory information database systems which would allow winery laboratory managers to answer basic performance, quality and cost efficiency questions. If such a system was to be created, it would need to be flexible enough to work in the wide range of sizes and requirements of our industry and simple enough to require little to no training of already overworked and undertrained staff.

Managing a winery lab requires at least some knowledge of chemistry, enology, compliance regulations, health and safety, and general administration. But even more valuable to managers would be the possession of a few basic facts that would assist them in directing their area more efficiently and economically. Unfortunately, these few basic facts are sometimes the most difficult ones to find without a high degree of effort, estimation or outright guessing. They are answers to these questions:

•   Exactly how many analyses of each type were performed last week/month/year?

•   What was the average cost of chemicals and supplies for each type of test?

•   What is the accuracy and precision performance data for each type of test?

•   How often are results from this type of test out of specification and how often must they be re-run?

How can a manager hope to improve lab efficiencies or manage a budget or staff without access to real answers to these fundamental questions? As basic as these questions may seem, very few winery labs can answer any of these questions without either laborious dissection of fussy old spreadsheets, going through thick binders and/or making educated or not-so-educated estimates.

Can you imagine your production manager not being able to say exactly how many cases of wine were bottling last week or last month? The manager of your barrel cellar counting every barrel by hand to know "for sure" how many barrels you have? What would you do if your vineyard manager had to make a guess at how many tons were harvested last year (and what if he is a new employee and didn't even work for you last year)? It would be considered intolerably poor business practices… so why do we tolerate it in so many of our winery labs?

Some History
Before microcomputers, lab results and calibrations were recorded manually and kept on clipboards and in binders. Traditionally, analysis was attached to wine lots by copying the results onto lot cards. To determine the total number of analyses was tedious but possible: someone would flip through the clipboard or binder and count them. Analyses would be written down twice--once on the analysis form and then copied once onto lot cards.

Now, with computers, lab results in most winery labs are still kept on clipboards and binders…but are also entered and copied on multiple Excel spreadsheets and located in various files, which are duplicated, mismanaged, modified, and impossible to sort out. Some, but not all, of this analysis data is then manually entered into bulk wine tracking software. I have lost count of the lab managers and technicians who must enter the same piece of data up to three times into various different programs or systems. Most of them also must copy by hand these data into other reporting forms. Tran­scription errors have become a major source of re-checks. Some data, especially the raw values and calibration data, are lost forever.

Stone Knives and Bearskins
This article was originally intended to discuss average industry values for the questions above, but it quickly became glaringly apparent that much more basic questions needed to be addressed. Instead, we will ask the question: Why don't winery labs have usable information systems appropriate for type of work, analysis and reporting performed in a winery lab?

Many of the labs visited and lab managers interviewed are fully aware of the limitations imposed on them by the lack of an appropriate data system. Many have stitched together inherited systems and antiquated reporting systems. Many have waited for years for the "new system," which turned out to be worse than the old one. Many have tried to generate reports using Excel spreadsheets formatted to imitate a true report while still maintaining the function of a spreadsheet and data entry system.

The way these multiple spreadsheets were set up means they cannot easily be linked, analyzed or collated. Mean­while, technicians enter some data into one spreadsheet, manually copy values to another and then, finally, enter a value into an official bulk tracking system. The data entered into the wine tracking system are forever separated from the relevant quality assurance data unless the trail is tediously unraveled by hand.

The Computer is Not a Typewriter
Many winery laboratories use the computer as a rather cumbersome and expensive typewriter, typing data into one table, printing it out, hand copying the values form one form to another. This increases both workload and frequency of transcription errors.

It is a basic concept of a database system that a value is entered one time. Computer systems should be set up such that once a piece of data is in the system, it never needs to be retyped or copied by human hands. If you evaluate your current system and find that you are typing the same data into a computer more than once--or worse, if you are hand-transcribing that data from one program to another--perhaps you would be better served by reverting to a paper system and just skipping the computer altogether. "Typing" into some computer program to allow the lab to print out labels for a list of wines that was generated by another computer is simply busywork and a waste of time and resources.

A Spreadsheet is Not a Database
The other extremely common frustration is the use of a spreadsheet as a poor substitute for a database. Spreadsheets are great for columnar reporting and manipulations but, for relational issues (such as linking calibration data to a set of results or tracking ongoing use of chemicals associated with a test), they are not appropriate. Possibly due to the ubiquitous nature of Excel to the exclusion of any other program, these nightmare spreadsheets have made their way from one winery to another, like some terrible flu virus, mutating slightly at each site before moving on (either from email to colleagues or by "taking it with you" when moving to another facility). Granted, with no other option, managers and technicians have tried to work with the "stone knives and bearskins" provided. Trying to go through nearly hundreds of Excel spreadsheets with inconsistent formatting, simply to tabulate a count of analyses and then estimate precision, is pretty unrealistic for any moderately occupied manager. The elaborate formatting used to make the spreadsheet appear like a report makes it difficult to archive data in a usable format which would allow simple evaluation of trends and values across multiple days and analytes.

A Lot Name is Not a Database Either
How much information must be included in a lot code? Apparently, from the lot codes existing in this industry, the code itself is the database. Vintage, vineyard, grower code, appellation, variety, harvest time, inoculation status, yeast type, barrel program, stability status, quality and price code are all database fields. Trying to fit all this information into the lot name is an accident waiting to happen (especially with any hand transcription or typing) and demonstrates the desperate steps wineries have taken to try to work within bad systems. In some cases, the lot name must be changed (manually) with every movement or treatment to reflect the new status.

A simple database would allow a lab to use reasonable lot names which make sense to a human--perhaps with basic vintage/varietal information and a unique identifying number--and the database would track all the other important factors. A good system would allow one to sort or find lots based on any parameter or number of parameters.

The Answer is Out There?
Laboratory Information Management systems (LIMs) are systems which integrate all the requisite lab processes (sample tracking, analysis requests, ordering, supplies, chemicals, quality control, quality assurance, coordination with equipment, reports, etc.) in one integrated location. They are common in medical and hospital labs, pharmaceutical companies, environmental labs and any facility with a professional, regulatory and/or financial interest in the quality of analytical results and convenience of a good management system.

Existing systems integrate with existing instrumentations (to receive sample lists and instructions and return results), generate control reports and charts and flag problems with performance, perform and record calibrations and calculations, integrate manual analysis requests and automatically generate them as instructed, print sample lists and labels, track supplies and chemical inventories, generate order lists, organize work lists and allow supervisors to manage the process and review and approve results while allowing all users to track progress and view/review results and reports.

Our basic management questions (total number of each type of analysis, cost per test, precision and accuracy, rework) could be answered by virtually any LIMs.

"There's no doubt there's a need for a lab information module for any bulk wine tracking database system," said Leighton Richardson at WMBD. "The question is if this is a commercially viable product."

Several commercial contract labs and at least one large California winery use the Perken Elmer program "Labworks." This system has modules and flexibility appropriate for large and multi-site facilities, capable of performing every standard LIMs function. However, it comes with a price tag starting in the range of $50,000, which is not likely to get a second glance from anyone familiar with the usual limitations of the average winery laboratory budget.

A Cheaper Answer, Perhaps?
A simple lab LIMs system would allow for just the basics and be flexible enough to be customizable for specific in-house issues. It should be able to integrate with the existing wine inventory system to the extent that data does not need to be re-entered but simply uploaded.

Most wine tracking systems need only three pieces of data to identify the correct location for analytical data: the date/time of the sample, the lot identity, and the container identity.

What would be some of the basic components of a winery laboratory information system? Certainly, tracking of supplies and chemicals, by date and lot number, price and volume, supplier and reorder timing, along with the ability to evaluate shelf life of chemicals and solutions. Next is the ability to run standard curves and calibrations and have them associated with a given set of results; in some cases values associated with these curves are used to generate the results; tied into this would be the quality control of the analyses, through some flexible control charting features, along with tracking systems for monitoring drift in calibrations, in standard curves, and in instrument performance.

Wide assortments of actions trigger the need for an analysis and each method must be integrated and manageable based on the individual winery. Analysis requests can be triggered by a physical process, a bookkeeping process, an addition, an analytical results, a predetermined time interval, a personal request or special circumstance, to name a few. The system should be able to easily generate reports to meet not just the winemaker's needs but also the needs of the lab manager, the technicians and the accountant. And finally, it must be able to answer simple questions like: How many analyses of each type were performed? What is the average cost of supplies and chemicals for each test? How much rework is needed for each type of test? What is the precision and accuracy of each test?

A Free Answer?
Last November, I spent some time working with a lab manager who was determined to design a "cheap" interim lab information system using Microsoft Access. He had coworkers with some Access experience and had used a nearly legendary in-house laboratory information Access database at another winery. He hoped to recreate what he had used previously and be able to share with his friends and coworkers.

After two days in discussion, much of it involving how to work around his company's existing lot code system (a complex code of 16 letters and numbers, which was scheduled to mutate as soon as the New System arrived), we parted ways. When next I spoke about his program, he was resigned to waiting for the new winery system--but dropped into my hands a copy of the mythical database…and it had just about everything the winery labs could use.

Now, if only we could learn to use Access the way we use Excel and then get permission to spread this particular database (or a customized version of it--or even one designed from scratch) around through email and migrating employees. wbm