Really Simple Systems

Organisations large and small depend on packaged software solutions for their core processes, and this is how it should be, the idea of writing your own ERP system or word processor makes no economic sense even if your business model is an extreme outlier. Differentiation does exist, the ERP system that a mining company would need would be significantly different from a chain of retail stores but most of the major (and minor) niches have packaged offerings to suit.

(Unless you’re a certain SME manufacturing company in the South-West of England that couldn’t find a suitable and affordable ERP package for its type of business; well then, developing an open-source one to help you and your fellow SMEs is, it seems, a viable option, resulting in uzERP!)

Many organisations will commission changes to these packages to better fit their way of doing business . In fact, such modifications are a major source of revenue for package providers (and as a result a major cost to those customers who will not “bend” to the packages’ default process flows).

The advice for those who either don’t have the money to commission changes (or who wish to avoid such costs) is to fit in with the package’s way of doing things (which as most will be “industry best practices” can often be a good thing) or use “off-line” processes to handle the exceptions. In some such cases a bespoke professionally delivered software solution might be the result; but for many the more usual “solution” to to those out-of-scope processes is “Simple!, you could easily do that in Excel“. Yet another recruit for the Excel -driven world of Shadow IT.

Whether such “systems” are the casualties of scope warfare or the preferred means of getting things done by Excel-savvy end-users they are part of every organisation’s IT landscape. They’re not going away, yet they receive scant attention from professional IT who often regard them as “somebody else’s problem”.

The end result is sub-optimal; although the business logic encoded in these spreadsheets is often of the highest quality and extremely fit-for-purpose (as it’s the work of those with the greatest insight of the problems at hand) the processes of moving data in and out (and when inside, efficently processing such data) is usually highly time-consuming, manual and error prone. This is where I come in.

I can wrap such “Simple Systems” is a professional “envelope” that will automate the flow of data in, out and within the process while keeping the likely-to-change business logic in Excel. Even when the Excel formula-based business logic needs to be converted to “code” (perhaps for performance or automation reasons) most likely that code will be in VBA (a language familiar to many citizen programmers) which can also be made configurable via a simple Excel “table” rather than some obscure XML dialect.

To get started, send me your existing workbooks (with or without data) with a short note describing the process they support and what you would like to automate or change. I’ll likely respond with some further questions re the process and perhaps made some further suggestions for automation. When we’ve agreed a “deliverable” I’ll get back to you with a fixed price quote.

The idea is to turn “Simple Systems” into “Really Simple System” not works of software art. The main interface will continue to be Excel enhanced by Excel-based tools (mainly my microETL add-in and perhaps add-ins such as PowerPivot for added reporting goodness). Separating data-flow from business logic is usually the major change. It’s also important to be clear what parts of the process need to remain “user-configurable” and which parts are to be “professionally” automated.

The user-configurable parts will need to continue to be Excel-formula based or as code that’s at a level the user is happy to engage with, or as code controlled via configuration tables. The professional parts will be developed by a professional (me!) with the expectation that a professional (i.e somebody proficient in Excel, VBA and SQL) would be required to modify the (always supplied) source code.

Highly configurable solutions are sometimes more expensive than “black boxed” ones, with the most user-configurable option – leave as is – the cheapest (until you take into account the payroll costs of those staff repeating the same easily-automated task over and over again!).

Contact Tom …


7 responses to “Really Simple Systems

  1. Tom

    Like the post – it spells out where most SMEs are in my experience…. i.e they have Sage or Quick Books (or something else) which ‘does the accounts’ and a bunch of unmanageable spreadsheets for everything else. Now substitute the SaaS accounting systems touted as the ‘next big thing’ in place of Sage/QB and nothing much will change – same old problems.

    The specific goal/motivation behind uzERP was to give SMEs (i.e. our SME in particular) an affordable way out of ‘spreadsheet-hell’ – it grew out of that into a more generic project so I guess in that sense we are in competition 😀

    However, there is a place for both the pre-canned (ERP) system and a more creative spreadsheet based model. Actually I’ve got a concept we’re building into uzERP called ‘user-space’ tables – you specify your table in our DB and share it for input/viewing etc. Pre-canned foreign keys (like a star schema I guess) to customer, stock etc will maintain some data-integrity and the fact that it is in the db means its secure, backed-up and shareable.

    Oh and using your ‘BI tool of choice’ i.e. excel, you can use ODBC to report/pivot etc to your heart’s content knowing the data is secure.

  2. Martyn,

    Not just an SME problem, my experience of such “systems” is in organisations running some of the most expensive and customised software on the planet. The sheer complexity of trying to centralise every sub-process and the fact that there’s only so many IT folks to go around always ensures a significant Shadow IT role. And although I call them “Simple Systems” many are anything but, particularly in areas such as pricing, budgeting & planning and managing and analysing external relationships and data.

    As with your idea of user-space tables, rather than trying to second-guess and deliver every single bit of business logic a better option for many IT organisations and package providers is to aim to become a reliable provisioner and distributor of clean and secure data.


  3. Tom,

    >>become a reliable provisioner and distributor of clean and secure data.<<
    Absolutely – get the data where it can be shared, analysed and generally used by as many people as poss. SMEs, though, usually have a deficiency in basic 'systems' and NO support and once they get beyond the accounting stuff things get VERY messy indeed.

    One company who are looking uzERP run all their planning on spreadsheets – looking at their process it should be in a central repository of some sort – they don't understand databases so couldn't really even build an access based system (ugh!!) and they can't afford the entry fee for an ERP package.

    So uzERP seems to fit the bill in this case.

  4. But yes, I agree – pricing, budgeting and forecasting are usually spreadsheet based and your Really Simple Systems approach, which ties stuff together, is about right.

  5. Pingback: Excel Document-Oriented Database with Python Map & SQL Reduce | Gobán Saor

  6. Pingback: Excel as a book of record. | Gobán Saor

  7. Pingback: When HAMMER met SWF | Gobán Saor