Friday Rant: Why Spend Analysis is Broken and How we Can Fix It — Forecasting and Beyond

I took a quick 24-hour swing to the Bay Area this week, cramming in four separate meetings with a number of old and new faces alike. In one of these meetings, I had the chance to drill into a relatively new spend analysis application in the market. The vendor in question asked me what I thought about the tool and what I'd do to improve it. Of course the usual things came to mind, but in the demonstration, the product manager used the word "forecasting" numerous times. And then it struck me. Why are spend analysis tools only rear-facing? What if we could actually incorporate elements of forecasting -- with varying confidence levels, ranges, etc. -- factoring in what spend might do versus simply gaining a rear-facing snapshot? For example, if we know spend is off contract or will soon be off contract and we know recent buying activity, why can't we forecast what the potential cost to the organization will be as a result?

It's not necessarily a simple calculation. In the above scenario, I'd need to know the amount bought off contracts in question and recent trending (adjusting for seasonality or one-time buys). And in the case of broader direct materials agreements, it would be useful to model the impact of underlying commodity price volatility on existing contracts and actual pricing paid relative to future pricing ranges based on current and implied volatility and different ranges. It might also be helpful to understand the types of agreements that are expiring based on the type of paper they're on and the clauses/clause levels contained (i.e., those with more favorable terms to suppliers vs. our organizations when it comes to payment terms, indemnification, non-solicits, etc.) Incidentally, there's software in the market today from providers like Seal Software whose contract extraction tools can gather and present such information on distributed contracts (i.e., on-servers, desktops) in an automated manner.

Taken together, this type of intelligence would help companies not just know that "100 contracts are expiring next month" but rather the 10-12 that they should really focus on to get the biggest bang for their spend buck based on their own specific objectives -- locking in cost savings, reducing commodity price volatility, increasing diversity spend, reducing risk, etc. With this type of information, spend analysis could truly become intelligent versus a static, periodically refreshed snapshot of our performance and trending to date. Yet I don't really see many vendors leaning in this direction over the short-term, let alone baking such concepts into a longer-term vision.

This disconnect between what we observe as a need for procurement and finance to better plan and forecast collaboratively in all areas of spend, category and commodity management and the current and emerging capabilities of spend analysis is suggestive of a market that is waiting for some breakthrough approaches. If you're a vendor and you have one, don't hesitate to let us know about it. Because based on the forecasting ability we know that leaders in the sector are looking for and that is wanting inside the available tools, we can largely describe spend analysis capabilities as broken. Or at the least built for a market of yesteryear -- and last month's snapshot of spending activity.

Jason Busch