There’s an interesting post over at Bruce Schnier’s blog where he discusses where security did, and didn’t, work with the Christmas underwear bomber incident. As is his usual inclination, he points out that the threat wasn’t new, security (on the whole) worked, and, of interest to us, the fact the more information would not have helped prevent the threat.
After the fact, it’s easy to point to the bits of evidence and claim that someone should have “connected the dots.” But before the fact, when there millions of dots – some important but the vast majority unimportant – uncovering plots is a lot harder.
This is a lot like the challenge we’ve been talking about under the banner of The value of information. How do we make sense of weak, conflicting and volumous signals we see in the environment outside our business, fuse this with strong signals from data inside the business, and create real insight? Granted, sometimes we’re aware of the signals (or at least the shape of their outline) we need to go looking for, much like Tesco’s decision to integrate weather forecasts and historical till information to predict customer demand. In other circumstances, we’re not so sure what we’re looking for. The business equivalent of predicting (and responding to) the underwear bomber might be managing exceptions in a complex, global supply chain, countering a competitor’s new product launch, or supporting a social case worker dealing with a unexpected crisis in a client’s domestic situation.
It’s tempting to create counter measures – prescriptive workflows designed to resolve a problem – to each of these scenarios on a case-by-base basis. Or even just throw up our hands and continue with the tribal processes of old. But, as Bruce points out, this doesn’t work. The challenge with taking action against specific threats is that the terrorist will simply use a new tactic next time, or you’ll be confronted with yet-another situation. Soon you’ll have overloaded your knowledge workers with exception scenarios which only address yesterday’s problems. You’ve started an arms race which you cannot win.
Bruce’s solution, in the context of security, is to integrate information into an operational decision making framework which wards against generic attacks.
What we need is security that’s effective even if we can’t guess the next plot: intelligence, investigation and emergency response.
This prompts me to think of two things:
First, we might need to add third dimension to that figure from Inside vs. Outside: Precision, to compliment Inside/Outside and Information Age. (Here, the engineer in me is going to split hairs over the definitions of focus, precise and accurate.) This new dimension captures how precise our need is. The Tesco example from above prefers precise signals, signal which communicates a single message. The exception manager might require imprecise signal, a derivative communicating a generic message aggregated generated by correlating a number of (in)precise signals. (A note of caution though, is to remember the recent impact of derivatives on the global financial markets.)
Second, we might want to rethink about how we conceptualise and use information information in our business. We currently have a very linear view, with information generation and consumption tightly connected to the stages of our value chain. It would be interesting to see how some of the ideas and frameworks behind the value of information could be fused with a decisioning framework like OODA. This would provide a tool to simplify the (potentially too complex) value of information framework, and realize it in operational work practices.
I’m not sure about the first point, but I expect the second will be fertile ground for further investigation.
Most interesting business decisions seem to be a synthesis process. We take a handful of data and fuse them to create an insight. The invention of breath strips is a case in point. We can rarely break our problem down to a single (computed) metric, the world just doesn’t work that way.
Most business decisions rest on small number of data points. It’s just one of our cognitive limits: our working memory is only large enough to hold (approximately) four things (concepts and/or data points) in our head at once. This is one reason that I think Andrew McAfee’s cut-down business case works so well; it works with our human limitations rather than against them.
I was watching an interesting talk the other day — Peter Norvig was providing some gentle suggestions on what features should be beneficial in a language to support scientific computing. Somewhere in the middle of the talk he mentioned the Curse of dimensionality, which is something I hadn’t thought of for a while. This is the problem caused by the exponential increase in volume associated with each additional dimension of (mathematical) space.
In terms of the problem we’re considering, this means that if you are looking for n insights to a problem in a field of data (the n best data points to drive our decision), then finding them becomes exponentially harder for each data set (dimension) we add. More isn’t necessarily better. While the addition of new data sets (such as sourcing data from social networks) enables us to create new correlations, we’re also forced to search an exponentially larger area to find them. It’s the law of diminishing returns.
Our inbuilt cognitive limit only complicates this. When we hit our cognitive limit — when n becomes as large as we can usefully use — any additional correlations can become a burden rather than a benefit. In today’s rich and varied information environment, the problem isn’t to consider more data, or to find more correlations, its to find the best three or features in the data which will drive our decision in the right direction.
How do we navigate from the outside in? From the decision we need, to the data that will drive it. This is the problem I hope the Value of Information discussion addresses.
I think we’re at a tipping point with BI. Yes, it makes sense that BI should be the next big thing in the new year, as many pundits are predicting, driven by the need to make sense of the massive volume of data we’re accumulated. However, I doubt that BI in its current form is up to the task.
As one of the CEOs Andy Mulholland spoke to mentioned “I want to know … when I need to focus in.” The CEO’s problem is not more data, but the right data. As Andy rightfully points out in an earlier blog post, we’ve been focused on harvesting the value from our internal, manufactured data, ignoring the latent potential in our unstructured data (let alone the unstructured data we can find outside the enterprise). The challenge is not to find more data, but the right data to drive the CEO’s decision on where to focus.
It’s amazing how little data you need to make an effective decision—if you have the right data. Andrew McAfee wrote a nice blog post a few years ago (The case against the business case is the closest I can find to it), pointing out that the mass of data we pile into a conventional business case just clouds the issues, creating long cause-and-effect chains that make it hard to come to an effective decision. His solution was the one page business case: capability delivered, (rough) business requirements, solution footprint, and (rough) costing. It might be one page, but there is enough information, the right information, to make an effective decision. I’ve used his approach ever since.
Current BI seems to be approaching the horse from the wrong direction, much like Andrew’s business case problem. We focus on sifting through all the information we have, trying to glean any trends and correlations which might be useful. This works as small to moderate scales, but once we reach the huge end of the scale it starts to groan under its own weight. It’s the law of diminishing returns—adding more information to the mix will only have a moderate benefit compared to the effort required to integrate and process it.
The information arms race, the race to search through more information for that golden ticket, is just a relic of the lack of information we’ve lived with in the past. In today’s land of plenty, more is not necessarily better. Finding the right features is our real challenge.
We’re drowning in a sea of data and ideas, with huge volumes of untapped information available both inside and outside our organization. There is so much information at our disposal that it’s hard to discern Arthur from Martha, let alone optimize the data set we’re using. How can we make sense of the chaos around us? How can we find the useful signals which will drive us to the next level of business performance, from amongst all this noise?
Traditional Business Intelligence (BI) tackles this problem by enabling us to mine for correlations in the data tucked away in our data warehouse. These correlations provide us with signals to help drive better decisions. Managing stock levels based on historical trends (Christmas rush, BBQs in summer …) is good, but connecting these trends to local demographic shifts is better.
Unfortunately this approach is inherently limited. Not matter how powerful your analytical tools, you can only find correlations within and between the data sets you have in the data warehouse, and this is only a small subset of the total data available to us. We can load additional data sets into the warehouse (such as demographic data bought from a research firm), but in a world awash with (potentially useful) data, the real challenge is deciding on which data sets to load, and not in finding the correlations once they are loaded.
What we really need is a tool to help scan across all available data sets and find the data which will provide the best signals to drive the outcome we’re looking for. An outside-in approach, working from the outcome we want to the data we need, rather than an inside-out approach, working from the data we have to the outcomes it might support. This will provide us with a repeatable method, a system, for finding the signals needed to drive us to the next level of performance, rather than the creative, hit-and-miss approach we currently use. Or, in geekier terms, a methodology which enables us to proactively manage our information portfolio and derive the greatest value from it.
I was doodling on the tram the other day, playing with the figure I created for the Inside vs. Outside post, when I had a thought. The figure was created as a heat map showing how the value of information is modulated by time (new vs. old) and distance (inside vs. outside). What if we used it the other way around? (Kind of obvious in hindsight, I know, but these things usually are.) We might use the figure to map from the type of outcome we’re trying to achieve back to the signals required to drive us to that outcome.
This addresses an interesting comment (in email) by a U.K. colleague of mine. (Jon, stand up and be counted.) As Andy Mulholland pointed out, the upper right represents weak confusing signals, while the lower left represents strong, coherent signals. Being a delivery guy, Jon’s first though was how to manage the dangers in excessively focusing on the upper right corner of the figure. Sweeping a plane’s wings forward increases its maneuverability, but at the cost of decreasing it’s stability. Relying too heavily on external, early signals can, in a similar fashion, could push an organization into a danger zone. If we want to use these types of these signals to drive crucial business decisions, then we need to understand the tipping point and balance the risks.
My tram-doodle was a simple thing, converting a heat map to a mud map. For a given business decision, such as planning tomorrow’s stock levels for a FMCG category, we can outline the required performance envelope on the figure. This outline shows us the sort of signals we should be looking for (inside good, outside bad), while the shape of the outlines provides us with an understanding (and way of balancing) the overall maneuverability and stability of the outcome the signals will support. More external predictive scope in the outline (i.e. more area inside the outline in the upper-right quadrant) will provide a more responsive outcome, but at the cost of less stability. Increasing internal scope will provide a more stable outcome, but at the cost of responsiveness. Less stability might translate to more (potentially unnecessary) logistics movements, while more stability would represent missed sales opportunities. (This all creates a little deja vu, with a strong feeling of computing Q values for non-linear control theory back in university, so I’ve started formalizing how to create and measure these outlines, as well as how to determine the relative weights of signals in each area of the map, but that’s another blog post.)
Given a performance outline we can go spelunking for signals which fit inside the outline.
Luckily the mud map provides us with guidance on where to look. An internal-historical signal is, by definition driven by historical data generated inside the organization. Past till data? An external-reactive signal is, by definition external and reactive. A short term (i.e. tomorrow’s) weather forecast, perhaps? Casting our net as widely as possible, we can gather all the signals which have the potential to drive us toward to the desired outcome.
Next, we balance the information portfolio for this decision, identifying the minimum set of signals required to drive the decision. We can do this by grouping the signals by type (internal-historical, …) and then charting them against cost and value. Cost is the acquisition cost, and might represent a commercial transaction (buying access to another organizations near-term weather forecast), the development and consulting effort required to create the data set (forming your own weather forecasting function), or a combination of the two, heavily influenced by an architectural view of the solution (as Rod outlined). Value is a measure of the potency and quality of the signal, which will be determined by existing BI analytics methodologies.
Plotting value against cost on a new chart creates a handy tool for finding the data sets to use. We want to pick from the lower right – high value but low cost.
It’s interesting to tie this back to the Tesco example. Global warming is making the weather more variable, resulting in unseasonable hot and cold spells. This was, in turn, driving short-term consumer demand in directions not predicted by existing planning models. These changes in demand represented cost, in the from of stock left on the shelves past it’s use-by date, or missed opportunities, by not being able to service the demand when and where it arises.
The solution was to expand the information footprint, pulling in more predictive signals from outside the business: changing the outline on the mud map to improve closed-loop performance. The decision to create an in-house weather bureau represents a straight forward cost-value trade-off in delivering an operational solution.
These two tools provide us with an interesting approach to tackling a number of challenges I’m seeing inside companies today. We’re a lot more externally driven now than we were even just a few years ago. The challenge is to identify customer problems we can solve and tie them back to what our organization does, rather than trying to conceive offerings in isolation and push them out into the market. These tools enable us to sketch the customer challenges (the decisions our customers need to make) and map them back to the portfolio of signals that we can (or might like to) provide to them. It’s outcome-centric, rather than asset-centric, which provides us with more freedom to be creative in how we approach the market, and has the potential to foster a more intimate approach to serving customer demand.
Tesco, the UK’s largest retailer, has started using weather forecasts to help determine what to stock in its stores across the UK.
Traditional approaches to stock management use historical buying data to drive stock decisions. This has worked well to date, but the increasing unpredictability of today’s weather patterns — driven by global warming — has presented business with both an opportunity and a challenge. An unexpected warm (or cold) spell can create unexpected spikes in demand which go unserviced, while existing stock is left on the shelves.
In Tesco’s own words:
In recent years, the unpredictability of the British summer — not to mention the unreliability of British weather forecasters — has caused a massive headache for those in the retail food business deciding exactly which foods to put out on shelves.
The present summer is a perfect example, with the weather changing almost daily and shoppers wanting barbecue and salad foods one day and winter food the next.
Tesco’s solution was to integrate detailed regional weather reports — valuable, external information — with the sales history at each Tesco store. A rise of 10C, for example, led to a 300% uplift in sales of barbecue meat and a 50% increase in sales of lettuce.
Integrating weather and sales data will enable Tesco to both capture these spikes in demand, while avoiding waste.
As Andy Mullholland pointed out in a recent post, all too often we manage our businesses by looking out the rear window to see where we’ve been, rather than looking forward to see where we’re going. How we use information too drive informed business decisions has a significant impact on our competitiveness.
I’ve made the point previously (which Andy built on) that not all information is of equal value. Success in today’s rapidly changing and uncertain business environment rests on our ability to make timely, appropriate and decisive action in response to new insights. Execution speed or organizational intelligence are not enough on their own: we need an intimate connection to the environment we operate in. Simply collecting more historical data will not solve the problem. If we want to look out the front window and see where we’re going, then we need to consider external market information, and not just internal historical information, or predictions derived from this information.
A little while ago I wrote about the value of information. My main point was that we tend to think of most information in one of two modes—either transactionally, with the information part of current business operations; or historically, when the information represents past business performance—where it’s more productive to think of an information age continuum.
Andy Mulholland posted an interesting build on this idea on the Capgemini CTO blog, adding the idea that information from our external environment provides mixed and weak signals, while internal, historical information provides focused and strong signals.
￼Andy’s major point was that traditional approaches to Business Intelligence (BI) focus on these strong, historical signals, which is much like driving a car by looking out the back window. While this works in a (relatively) unchanging environment (if the road was curving right, then keep turning right), it’s less useful in a rapidly changing environment as we won’t see the unexpected speed bump until we hit it. As Andy commented:
Unfortunately stability and lack of change are two elements that are conspicuously lacking in the global markets of today. Added to which, social and technology changes are creating new ideas, waves, and markets – almost overnight in some cases. These are the ‘opportunities’ to achieve ‘stretch targets’, or even to adjust positioning and the current business plan and budget. But the information is difficult to understand and use, as it is comprised of ‘mixed and weak signals’. As an example, we can look to what signals did the rise of the iPod and iTunes send to the music industry. There were definite signals in the market that change was occurring, but the BI of the music industry was monitoring its sales of CDs and didn’t react until these were impacted, by which point it was probably too late. Too late meaning the market had chosen to change and the new arrival had the strength to fight off the late actions of the previous established players.
We’ve become quite sophisticated at looking out the back window to manage moving forward. A whole class of enterprise applications, Enterprise Performance Management (EPM), has been created to harvest and analyze this data, aligning it with enterprise strategies and targets. With our own quants, we can create sophisticated models of our business, market, competitors and clients to predict where they’ll go next.
Despite EPM’s impressive theories and product sheets, it cannot, on its own, help us leverage these new market opportunities. These tools simply cannot predict where the speed bumps in the market, no matter how sophisticated they are.
There’s a simple thought experiment economists use to show the inherent limitations in using mathematical models to simulate the market. (A topical subject given the recent global financial crisis.) Imagine, for a moment, that you have a perfect model of the market; you can predict when and where the market will move with startling accuracy. However, as Sun likes to point out, statistically, the smartest people in your field do not work for your company; the resources in the general market are too big when compared to your company. If you have a perfect model, then you must assume that your competitors also have a perfect model. Assuming you’ll both use these models as triggers for action, you’ll both act earlier, and in possibly the same way, changing the state of the market. The fact that you’ve invented a tool to predicts the speed bumps causes the speed bumps to move. Scary!
Enterprise Performance Management is firmly in the grasp of the law of diminishing returns. Once you have the critical mass of data required to create a reasonable prediction, collecting additional data will have a negligible impact on the quality of this prediction. The harder your quants work, the more sophisticated your models, the larger the volume of data you collect and trawl, the lower the incremental impact will be on your business.
Andy’s point is a big one. It’s not possible to accurately predict future market disruptions with on historical data alone. Real insight is dependent on data sourced from outside the organization, not inside. This is not to diminish the important role BI and EPM play in modern business management, but to highlight that we need to look outside the organization if we are to deliver the next step change in performance.
Zara, a fashion retailer, is an interesting example of this. Rather than attempt to predict or create demand on a seasonal fashion cycle, and deliver product appropriately (an internally driven approach), Zara tracks customer preferences and trends as they happen in the stores and tries to deliver an appropriate design as rapidly as possible (an externally driven approach). This approach has made Zara the most profitable arm of Inditex, a holding company of eight retail brands, and one of the biggest success stories in Spanish business. You could say that Quants are out, and Blink is in.
At this point we can return to my original goal: creating a simple graphic that captures and communicates what drives the value of information. Building on both my own and Andy’s ideas we can create a new chart. This chart needs to capture how the value of information is effected by age, as well as the impact of externally vs. internally sourced. Using these two factors as dimensions, we can create a heat map capturing information value, as shown below.￼
Vertically we have the divide between inside and outside: internally created from processes; though information at the surface of our organization, sourced from current customers and partners; to information sourced from the general market and environment outside the organization. Horizontally we have information age, from information we obtain proactively (we think that customer might want a product), through reactively (the customer has indicated that they want a product) to historical (we sold a product to a customer). Highest value, in the top right corner, represents the external market disruption that we can tap into. Lowest value (though still important) represents an internal transactional processes.
As an acid test, I’ve plotted some of the case studies mentioned in to the conversation so far on a copy of this diagram.
The maintenance story I used in my original post. Internal, historical data lets us do predictive maintenance on equipment, while external data enables us to maintain just before (detected) failure. Note: This also applies tasks like vegetation management (trimming trees to avoid power lines), as real time data and be used to determine where vegetation is a problem, rather than simply eyeballing the entire power network.
The Walkman and iPod examples from Andy’s follow-up post. Check out Snake Coffee for a discussion on how information driven the evolution of the Walkman.
The example from my follow-up (of Andy’s follow-up), of Albert Heijn (a Dutch Supermarket group) lifting the pricing of ice cream and certain drinks when the temperature goes above 25° C.
Netflix vs. (traditional) Blockbuster (via. Nigel Walsh in the comments), where Netflix helps you maintain a list of files you would like to see, rather than a more traditional brick-and-morter store which reacts to your desire to see a film.
Send me any examples that you know of (or think of) and I’ll add them to the acid test chart.
An interesting exercise left to the reader is to map Peter Drucker’s Seven Drivers for change onto the same figure.
Andy Mulholland has a nice build on my value of information bit over at Capgemini’s CTO blog, flipping the sense of the figure and showing how the time axis also connects to internal vs. external focus, and IT’s shift from cost control to value creation.
Update 2: Andy Mulholland came across a nice example:
Albert Heijn the Dutch Supermarket group lifts the pricing of ice cream and certain drinks when the temperature goes above 25’ C
Update 1: I’ve left a comment there building on what Andy has.
BI does seem to be moving in this direction, but still has a long way to go and is too internally focused. Customer Intelligence is moving the enterprise boundary out a little, and does not really address the challenge of integrating external information to create new insight. What about local events, weather, the memes from the social media community, the memes from our competitors customers, or anything else we can think of? The challenge is to fuse internal, customer, competitor, market and even environmental data to create new insight.
For example, consider current approaches to S&OP (sales and operations planning). We’ve take what is an inherently unstructured and collaborative activity and shoved it through the process and business intelligence meat grinder to create yet-another enterprise application. It’s no surprise that S&OP is a challenge to deploy, with few companies realizing (let alone capturing) the promised value. Customer Intelligence adds little to the benefit side of this this equation; it would seem impossible to justify CI in terms of cost saving, and challenging to justify it in terms of creating new business.
Imaging a world where we have our S&OP team focused on information synthesis rather than the planning process. They might pluck weather data (it’s going to be hot in St Kilda) and couple it with an event (the St Kilda festival), memes from their customers (and their competitor’s customers) plucked from hootsuite, and decide only 24 hours before the event to rapidly deploy a pop-up store. It’s this sort of sense-and-respond ability that will drive us to the next level of performance.
One of the best real world examples of this transition from internal-cost-control to external-value-capture has happened around the hand-held stock management devices used in retail. Initial deployed as a cost control measure (i.e. better information on what’s on the shelves) they have now become a tool for capturing value. Walmart has been using these devices for some time, devolving buying decisions to the team walking the shop floor and providing them with the information they need to make good buying decisions. As one reporter found:
“We received an inspirational talk on this subject, from an employee who reacted after the store test-marketed tents that could protect cars for people who didn’t have enough garage space. They sold out quickly, and several customers came in asking for more. Clearly this was a singular, exceptional case of word-of-mouth, so he ordered literally a truckload of tent-garages, “Which I shouldn’t have done really without asking someone,” he said with a shrug, “because I hadn’t been working at the store for long.” But the item was a huge success. His VPI was the biggest in store history—and that kind of thing doesn’t go unnoticed in Arkansas.”
In BI terms, we’re moving from large, centralized solutions used to drive planning, to distributed peer-to-peer networks focused on supporting local decisions. While corporate data stores will still play an important role, the advantage is moving to our ability to fuse multiple data sources, some which we do not own and some which only have local relevance. The right information, at the right time, in the right place, to empower knowledge workers to make the best possible decisions. Local Intelligence, rather than Business Intelligence.
We all know that data is valuable; without it it would be somewhat difficult to bill customers and stay in business. Some companies have accumulated masses of data in a data warehouse which they’ve used to drive organizational efficiencies or performance improvements. But do we ever ask ourselves when is the data most valuable?
Billing is important, but if we get the data earlier then we might be able to deal with a problem—a business exception—more efficiently. Resolving a short pick, for example, before the customer notices. Or perhaps even predicting a stock-out. And in the current hyper-competitive business environment where everyone is good, having data and the insight that comes with it just a little bit sooner might be enough to give us an edge.
A good friend of mine often talks about the value of information in a meter. This makes more sense when you know that he’s a utility/energy guru who’s up to his elbows in the U.S. smart metering roll out. Information is a useful thing when you’re putting together systems to manage distributed networks of assets worth billions of dollars. While the data will still be used to drive billing in the end, the sooner we receive the data the more we can do with it.
One of the factors driving the configuration of smart meter networks is the potential uses for the information the meters will generate. A simple approach is to view smart meters as a way to reduce the cost of meter reading; have meters automatically phone readings home rather than drive past each customer’s premisses in a truck and eyeball each meter. We might even used this reduced cost to read the meters more frequently, shrinking our billing cycle, and the revenue outstanding with it. However, the information we’re working from will still be months, or even quarters, old.
If we’re smart (and our meter has the right instrumentation) then we will know exactly which and how many houses have been affected by a fault. Vegetation management (tree trimming) could become proactive by analyzing electrical noise on the power lines that the smart meters can see, and determine where along a power line we need to trim the trees. This lets us go directly to where work needs to be done, rather than driving past every every power line on a schedule—a significant cost and time saving, not to mention an opportunity to engage customers more closely and service them better.
If our information is a bit younger (days or weeks rather than months) then we can use it too schedule just-in-time maintenance. The same meters can watch for power fluctuations coming out of transformers, motors and so on, looking for the tell tail signs of imminent failure. Teams rush out and replace the asset just before it fails, rather than working to a program of scheduled maintenance (maintenance which might be causing some of the failures).
When the information is only minutes old we can consider demand shaping. By turning off hot water heaters and letting them coast we can avoid spinning up more generators.
If we get at or below seconds we can start using the information for load balancing across the network, managing faults and responding to disasters.
I think we, outside the energy industry, are missing a trick. We tend to use a narrow, operational view of the information we can derive from our IT estate. Data is either considered transactional or historical; we’re either using it in an active transaction or we’re using it to generate reports well after the event. We typically don’t consider what other uses we might put the information to if it were available in shorter time frames.
I like to think of information availability in terms of a time continuum, rather than a simple transactional/historical split. The earlier we use the information, the more potential value we can wring from it.
There’s no end of useful purposes we can turn our information too between the billing and transactional timeframes. Operational excellence and business intelligence allow us to tune business processes to follow monthly or seasonal cycles. Sales and logistics are tuned on a weekly basis to adjust for the dynamics of the current holiday. Days old information would allow us to respond in days, calling a client when we haven’t received their regular order (a non-event). Operations can use hours old information for capacity planning, watching for something trending in the wrong direction and responding before everything falls overs.
If we can use trending data—predicting stock-outs and watching trends in real time—then we can identify opportunities or head off business exceptions before they become exceptional. BAM (business activity monitoring) and real-time data warehouses take on new meaning when viewed in this light.
In a world where we are all good, being smart about the information we can harvest from our business environment (both inside and outside our organization) has the potential to make us exceptional.