One-piece Flow for Information Products?

Organizations that produce documents — whether they are publications for sale, standard tests for schools, legal templates, or work instructions for production — face challenges that differ from manufacturing, because data and materials don’t flow the same way. The production of a document by a team is a process of collaborative editing, not a fixed sequence of standardized operations.

With electronic documents, you need a revision management system to prevent inconsistent updates, you need to cap the number of documents in process to control lead time, and you may need to improve the work flow or increase the team size if saturated.

Tools like 5S are irrelevant in this context, because the work takes place inside a computer network, not in the physical office, and setting up an effective network — with the right software properly configured — requires information systems professionals at the state of the art. What looks like rework in this context is a collaborative editing process that must be managed, not eliminated.

Document Generation Versus Manufacturing

First,  asked the TPS Principles and Practice discussion group on LinkedIn for advice on helping a client move “from batch-and-queue to one-piece flow,” with an explanation that suggested it was in the manufacture of custom products, and the members duly responded. But Edythe is not involved with manufacturing. She works in education management, and her expertise instead is in “selection and test development/validation, research, assessment, technical report writing, documentation, statistics, databases.”

Later in the discussion, she clarified the context by saying “the product is a unit of information, not a physical piece. Each is unique and there is much redesign during inspection.” Even after she said that, some members argued that it doesn’t matter what you build it is about how you build it, and recommended starting with 5S.

That manufacturing work pieces are physical objects is not a detail. Materials and data don’t flow the same way. Once you pass a work piece to the next operation in a production line, you no longer have it. Data, on the other hand, is passed on like a contagious disease. After you have passed it to the next person, you still have it.

The key issue in a group of people producing a document together is to prevent inconsistent updates. If you pass a paper document around, it can only be modified by whoever physically has it. Electronic documents, on the other hand, are plagued by divergent, inconsistent updates on parallel copies.

To avoid this, you need the document in a database rather than have  copies of it floating around as email attachments. And you need a revision management system and a workflow that grants exclusive update access for each documents to contributors in a prescribed sequence, with others possibly allowed to mark them up with comments without modifying them.

It is a different challenge. You have to worry about these issues but not about stock-outs or parts presentation. Manufacturing organizations have this challenge too, in the management of their metadata — product nomenclature, process specifications, work instructions, etc.

Organizing the information system of a company for such purposes is a job for pros at collaborative software and database or data warehouse design. If you entrust it to holdovers from the paper-and-pencil era, they will do counterproductive things like standardizing file names with some “smart” numbering system instead of using descriptive names. And they won’t know any better than using “simple” tools like Excel as data stores, or distributing documents as email attachments. Not to mention the countless hours that will be spent training others to comply with archaic “standards.”

Even the best workflow is subject to saturation if you shove too much work into it. If you want anything resembling one-piece flow in production, you need to cap the number of documents in process. You can do it even if the amount of work varies with each document.

This will keep the processing lead time short, but not the order fulfillment lead time, because documents will still wait upfront. The benefit you will get is that errors will be detected and corrected faster and better.  Then, if you need to reduce the order fulfillment lead time, you will need to increase productivity — by improving each step in the workflow and balancing the steps — or add capacity.

Irrelevance of 5S in Document Generation

One of the respondents said “it doesn’t matter what you build, it is about how you build it,”  and went on to recommend starting with implementing 5S. In reality, what you make, how often you make it, how much variety in mix and volume you have, and how capable you are at making it, … all matter a great deal.

If all the work for Edythe’s products is being done inside a computer network, what exactly do you expect 5S on employees’ physical desktops to accomplish? Their paper inboxes may be neat and their pencils in a marked location. Most office 5S efforts also extend to attendance boards with magnets, bookshelves, bulletin boards, etc. So what? It’s the organization of their hard drives that makes a difference.

And when Kaizen efforts reach into information systems, it is often to do more harm than good, for example by deleting historical data, when a well-organized on-line archive is often much more valuable than the disk space freed by deleting it.

Rework Versus Collaborative Editing

Edythe is describing the collaborative editing process for a document as rework, but rework is what you do to fix a workpiece that is branded defective based on consistent criteria through a test, an inspection, or a gauge check.

A document prepared by a team has to be iteratively refined in order to include the wisdom of all its co-authors. It’s not rework as we know it in manufacturing, and it’s an activity that needs to be organized, not eliminated.