For most large businesses, rolling out a major initiative company-wide is always high-cost. Many companies plough ahead based on gut feel or a degree of manual financial analysis. So much more, however, can be gained through the rigorous business testing which is now possible using advanced technology, as Phil Marsland, European head of Applied Predictive Technologies, explains.
When rolling out a major initiative, the stakes are particularly high for large, consumer-focused companies. If you run a retail company, a decision to refurbish all your outlets could cost you millions. The same applies if you are a hotel group planning to open in a series of new locations, or a bank intending to extend your opening hours.
Of course, major investments are very worthwhile if they lead to greatly increased profit and better customer satisfaction. The gains could well, however, be insufficient to offset the costs of implementation. In the worst cases, they could lead to disastrous, unexpected losses.
So how can company directors accurately predict the outcome of proposed investments?
Some businesses take such major decisions based simply on gut feel. Others adopt a more thorough approach, formulating a hypothesis and conducting a trial. A retailer might try a marketing promotion, for example, in a small number of stores; or a bank may change the layout of a few branches. The company will monitor the trial and a team of financial analysts might pore over spreadsheets of data before presenting the findings to the board—with a recommendation for or against wider roll-out.
That might sound good, but here is the bad news: this approach is typically too flawed to be of significant value. A trial such as this will usually provide some valuable operational lessons. In terms of providing accurate insight into key performance metrics, however, such a trial might even be worse than no trial at all, as misleading findings can give a company false confidence to implement a change that will harm the business—or to withhold an initiative that would have been of great advantage.
The fact is, accurate small-scale trials are exceptionally difficult to set up and run in a statistically robust manner, since the findings can so easily be distorted or completely masked by all the surrounding ‘noise’. For retailers, for example, this might include the weather, local transport issues and competitor activity around the stores taking part in the trial of a new promotion.
The good news is that the vast computing power now available to companies means that, with the right tools, you can cut through all this noise and conduct fully meaningful business trials, even when working with small samples. Such an approach will ensure a trial is scientifically set up, with carefully chosen test and control groups, while accounting for all the misleading factors that would otherwise obscure the result. Not only will this provide an accurate ‘yes/no’ answer to whether an initiative will benefit the business, but a disciplined process of ‘test and learn’ will also help a company analyse exactly how the initiative can be tailored for optimal effect.
A hotel group could learn, for example, that refurbishing outlets used by business people will increase profits; while doing the same for outlets where leisure visitors stay will only increase costs. A restaurant chain may learn that increasing waiting staff will boost takings in city centres, but not out of town. So companies can target proposed initiatives for maximum benefit based on certain products, geographical areas, customer profiles and so on.
There are so many initiatives that can be tested in this way—right across the spectrum of marketing, capital and operational investments—and the need to do so has been made all the more acute by the pressures of the economic downturn. Shoppers, for instance, are increasingly filling their baskets with products that are on offer. What are the most effective promotions that will make the smallest dent in retailers’ profits? Testing options scientifically will provide an answer.
Economic pressures mean companies are wrestling with the need to minimise operational costs. But which areas of expenditure can be safely trimmed and which would harm the business? Rigorous testing can help provide an answer.
Needless to say, this approach can bring huge benefits. Our experience over the years has shown that, by institutionalising a ‘test and learn’ approach organisation-wide, a company can increase its profit margins by up to five per cent.
Having worked across the world, my company has found that North American companies are generally ahead of the rest, with many having used advanced testing techniques for a considerable time. We have been privileged to work with the likes of Starbucks, Subway, InterContinental Hotels and Wells Fargo, many of whom have embedded rigorous testing deeply into their company cultures.
Some examples may help to bring this to life. The US convenience store group Wawa planned to introduce a new breakfast product onto its shelves. The company admits that it used to ‘take chances’ with such decisions based on its best manual interpretation of data. In this case, the breakfast item seemed a good product; it received a positive reception in spot tests; and the management team was poised for company-wide roll-out.
At the time, however, the company was trying a new software-based approach to business testing, and assessing the proposed breakfast product using this system revealed a very different story. It showed what had previously gone unnoticed: that it would cannibalise sales of higher margin products and lead to a net dent in overall profits. Needless to say, the company killed off the product launch.
Big Lots is another US grocery group which has adopted advanced systems for business testing. Among many other initiatives, it tested the number of printed circulars distributed to its stores. It has consequently saved several million dollars per year, as the tests showed that circular numbers could be significantly reduced without harming sales. Big Lots also trialled alterations to staff hours, enabling it to work out how—and in which stores—it could reduce staffing in response to the downturn without reducing takings. The company has embraced a culture of test and learn throughout its business and says this has helped it achieve 14 consecutive record quarters, despite a tough overall retail environment.
Our experience that other regions generally lag North American practice was underlined by a recent UK study by analysts Martec International. Martec surveyed large retailers representing over half the UK sector and found that fewer than a third are ‘very confident’ about their ability to trial investments. (Based upon our observations from client relationships, I would suspect that many of even this third could be doing it more effectively). That said, this situation is starting to change and the first few innovative companies outside North America are waking up to the benefits of a scientific test and learn approach.
In Taiwan, retailer Family Mart was debating whether to install coffee machines in its stores. Would this increase overall sales or would it detract from purchases of other drinks? Informally testing in a few stores would have produced misleading results because of the myriad of factors at play; whereas installing machines in all 2,480 stores to see what happened would have been a wild leap of faith.
The company resolved its dilemma by undertaking a trial, using the latest testing software, based on carefully selected test and control stores. While some outlets saw benefits, others saw none, and this enabled Family Mart to draw up a prioritised list of stores that should receive machines based on the trial data generated.
In the UK, we have recently been working with Boots, the health and beauty retailer. The company first applied advanced testing tools to trials of a store refurbishment programme. This enabled it to analyse which elements of the programme would be most productive and how quickly it could expect to recover its investment. Boots is now using this approach right across its operation with many diverse tests underway.
Similarly, UK high street jewellery retailers H. Samuel and Ernest Jones have just embarked on an equally rigorous approach, despite being much smaller companies than Boots. They are using testing technology to trial marketing initiatives, capital investments such as store changes and operational matters such as alterations to staffing, as well as gaining greater insight into shopper behaviour and the impact of promotional schemes.
So the pitfalls for large companies of running insufficiently rigorous trials can be huge—and all the more damaging in this tough trading climate. Of course, embedding a truly scientific test and learn culture company-wide requires considerable investment itself: in technology, in training and not least in senior management focus. The benefits, however, are generally overwhelming.
With a potential five per cent increase in profitability at stake, few large companies can afford to be doing ‘business as usual’.
Phil Marsland is European head of Applied Predictive Technologies, which specialises in helping companies implement rigorous business testing methods.
