(This is Part 2 of an article on a topic of some concern to most SME businesses. It can be read as a continuation of Part 1, or as a stand-alone piece. Click here to read Part 1.)

Everyone needs to adjust their mental model. Computerization efforts are more likely to succeed if their implementers recognize and address a quite subtle problem, which is, that it's really hard to escape the clutches of an obsolete mental model. It's a subtle problem because we don't often question our mental models; in fact, we tend to see everything through the filter of our current paradigms, whether they're still valid or not. It takes quite a jolt to get us to change the way we view the world.

Let's examine the work paradigm in non-computerized organizations.

In non-computerized organizations, people not only do their main tasks – like sales, or purchasing, or cashiering – they also document and record what they have done. Everybody in a paper-based system is both a doer of tasks and a documenter recording what he has done. Sales departments negotiate with prospective customers and prepare sales reports; purchasing departments negotiate prices and specs with suppliers and prepare purchasing reports. Warehouse clerks receive and issue inventory and prepare inventory reports.

Since doing-and-recording has been going on ever since mankind invented writing, not surprisingly most people see this as the natural state of things.

But in a paper-based system, preparing reports usually takes much longer than the underlying tasks themselves; so organizations respond by batching their reporting. Salesmen will close sales and accumulate invoices all week, and then prepare a sales report on Friday. Warehouse people dutifully log inventory ins and outs, then prepare a stock-on-hand report on Friday. Similarly, Purchasing will issue purchase orders all week, then prepare a purchase report at week's end.

But real life is messy, and not all transactions will unfold smoothly. Some snag or other always pops up to delay transactions, and always for seemingly good reasons - a question about a point of fact, or a point of control, or an issue of funding. And so, a system that's already slow by nature... gets even slower.

To the insightful CEO or COO watching this cycle month in and month out, the bottom line is that things in a manual system move too slowly and are too prone to defects (even as colleagues who don't have the big-picture perspective continue to view these delays and defects as “unavoidable realities” or “business as usual”).

When an organization realizes that its slow, herky-jerky back room is hampering its growth, or hurting customer satisfaction – that's when it starts thinking about computerizing.

Now, one of the most alluring promises of computerization is that the age-old pattern of doing-and-recording will get streamlined. The computer makes the recording part of doing-and-recording supremely fast and effortless, so that the day-to-day work reality for most workers is transformed. 

Sales people can now spend a lot more of their time with customers,  and far less of it shuffling paper, thus creating happier customers. Credit and collections folks can now spend more time doing credit investigation and billing customers; they can now spend zero time cobbling up receivables ageing reports or billing statements, because the computer does all that automatically. Everybody is now able to spend more of their workday doing real things in the physical world (and being productive), and far less non-revenue time making reports.

Now, a big obstacle to a successful changeover is that our thought processes tend to cling to the old manual mental model even as we already have information technology solutions in place. This is not really surprising; people will instinctively continue to apply old patterns of thinking to new realities... until they learn better. So, people who unwittingly operate under an obsolete mental model will continue to batch-process their work, just like they've done all their working lives. This is the danger point for the transition leader. This is when managers are in danger of losing control of the computerization effort.

What's the danger here? First, recognize that most businesses operate in a circular cycle; in a trading business it's: purchase stocks, receive supplier deliveries, sell merchandise,  then purchase stocks again, and so on. Therefore, everyone's output is needed by someone further down the cycle. When users input data at their own convenience, according to their own self-established deadlines, they in effect ignore downstream co-workers and processes that rely on their output.

For example, if sales quantities are not being inputted in time, real inventory in the warehouse will not jibe with stock on hand per the computer's records (the merchandise is already gone – sold – but the computer reflects that it's still in the warehouse). Because inventory is thus overstated in the reports, Purchasing will order insufficient quantities during the next buying cycle. And because the branches then end up getting too little inventory, stockouts occur and the company loses sales.

Let this happen a few times, and people all over the company start losing faith in the computer data. (“How can we believe the computer when it causes stock-outs? Something's wrong with the software!”) And when people no longer believe the computer, they start instinctively creating compensating mechanisms. They start creating freestanding islands of departmental data. Pretty soon the main departments are offering numbers from their own data hoard that competes with the computer data. This makes things worse, not better.

Your problem is that now you have two or more sources of information; which one will you believe? At this point, you have lost the clarity that comes from having a single source of truth,  because you lost your nerve at the first sign of trouble and allowed that second and third source of competing truth into the picture.  Isn't this precisely the scourge of a paper- based system? At that point, the computerization effort has fallen back into the slimy embrace of a zombie resurrection of the old manual system. And at that point, you can conclude that the computerization project is in serious trouble.

Sound familiar?

How to avoid this fate? The implementers of computerization must internalize and communicate the following basic ideas to everyone in the company:

1. Maintaining the computer data must be subject to time discipline – specifically, strict intra-day deadlines for data-encoding. Workers should not be allowed to unilaterally choose when they will input data; instead, they should be required to input transactions within, say, 30 minutes of the occurrence of the underlying business event. Only when data is inputted in a disciplined, time-conscious way, can the data be counted on to be up-to-date, and therefore believable and reliable.

For example, in one retail organization we consult for, sales from the branch point-of-sale systems must be uploaded to the main back-room system by 6 pm everyday. This assures everyone, especially other departments, that at 6 pm, they can have a picture of inventory that they can count on and base decisions on. If there were no 6pm deadline, if the sales analyst could upload POS data anytime he pleased, nobody would know when they could get a reliable inventory figure. Does this inventory number already reflect the day's sales or not?

Without time discipline, decisions that need a hard inventory figure – say, purchasing – would require a delicate and time-consuming back-and-forth between the parties involved before useful actions could be enacted. With a judiciously established deadline, there is clarity, there is no fuss, and the computer data can be counted on for serious decision-making. With high confidence in the completeness of the stock-on-hand data, the purchasing department can in turn start processing purchase orders by 6pm. It can then promise that all PO's will be out by, say, 12 noon of the following day.

3) Do not allow competing islands of information to coexist with the computer data. When users express skepticism about the data in the computer, fix the basic problem by addressing the cause; do not lose your nerve and give in to the easy but false solution of parallel data. In the above example, the way to eliminate vague data was to impose clear cut-off times (deadlines) on data inputting, and not to create supplementary freestanding data. The proper response was to make the single source of truth more reliable, not to allow competing and potentially conflicting sources of data into the company's information flow.

4) One sure way to doom a computerization effort is to delegate it to middle management. Senior executives with hidden agendas – or who merely lack enthusiasm – can easily undermine middle-manager types. To forestall this, the CEO or COO must be personally engaged with the computerization rollout. Nobody expects the top guy to be a computer expert; that's not what he's needed for. What he's needed for is, first, his deep understanding of the business, and second, the clout to knock (hard) heads together. Nobody outranks the top guy; so the project can't be stonewalled by other senior types. 

Being engaged includes (a) showing through word and deed management's strong commitment to the computerization project, and (b) bringing to bear sufficient management authority to resolve certain disputes that are sure to arise in the course of the project. Computerizing a company is often a tricky, contentious process. Some people will oppose it out of fear of change, fear of loss of power, or from sincere convictions about alternative ways of doing things. Top management is the only entity strong enough to get the last word, and make it stick. RSR

back to top >>>