Chapter 2

The ACCC's GROCERYchoice website

Launch of the GROCERYchoice website

2.1        The GROCERYchoice website[1] was launched by the Australian Competition and Consumer Commission (ACCC) on 6 August 2008, the day after the release of the ACCC's report on its inquiry into the competitiveness of retail prices for standard groceries. 

2.2        The ACCC report had found that while the Australian grocery market was 'workably competitive', consumers would benefit from more vigorous competition.  It concluded that there were high barriers to entry and expansion in the grocery market, limited incentives for the major supermarket chains to compete aggressively on price, and limited price competition to the major chains from the independent sector.  The entry of ALDI in the eastern states had provided a substantial beneficial impact on competition.[2] 

2.3        Announcing the launch of the website, the then Minister for Competition Policy and Consumer Affairs, the Hon. Chris Bowen MP, said:

It will provide a monthly snapshot of grocery prices in 61 different baskets – meat and seafood, fruit and vegetables, dairy, bread and cereals, drinks and snacks, general groceries, household and personal care; and basic staples basket to allow comparisons with ALDI, who do not have a wide enough range of goods to be compared with the other retailers on the different baskets.

The ACCC has undertaken considerable work to ensure that only like for like goods are used for comparison purposes for these baskets.  The goods in the baskets will change from month to month, and be kept confidential to ensure supermarkets cannot manipulate prices to artificially reduce their prices in that basket while putting up prices on other goods.[3]

2.4        The Minister noted some 'interesting facts' to come out of the first month of surveys, including that Coles was the cheapest of the major supermarkets in 52 out of 61 regions.  In the 40 out of 61 regions where ALDI was present, that retailer had the cheapest prices on basic staple products, with a difference of around 25 per cent compared to the next cheapest retailer. 

2.5        Mr Bowen also suggested that the introduction of the website could change consumers' shopping patterns, encouraging them to shop elsewhere, rather than their 'usual' supermarket:

Well, what [the website] does is give people a guide to say, well, in our region, ALDI, for example, is significantly cheaper. So people get into a habit. I'm no different and I'm sure people in this room are no different.  You get into a habit. You go to the same supermarket time and time again and you notice prices going up and you think that's not good. 

But it's not a spur for you often to look around, to search for cheaper prices.  And if you do, you may get there and find that the prices are no different.  So this is a guide for people to say, well, in your region, if choose to have a look at Woolworths, or Coles, or ALDI, or independents, whatever the case may be in that region, you may find a difference in prices. And then those people, as they should in a market economy, can make their own choices based on convenience, customer service, quality, et cetera. 

2.6        The Chairman of the ACCC, Mr Graeme Samuel, commented at the press conference that GROCERYchoice would provide new information to consumers:

[GROCERYchoice] won't duplicate supermarket advertising; it won't publish the weekly specials; it won't list grocery prices for individual supermarkets. But this is about giving consumers something new that the supermarket chains won't tell them: who is cheapest in each region overall.[4]

2.7        The website's introductory page provided the following information:

GROCERYchoice ... provides practical grocery price information to help consumers find the cheapest overall supermarket chain in their area. With the large number of grocery items available at each supermarket, consumers often find it difficult to determine which retailer offers the cheapest prices overall. GROCERYchoice helps consumers compare the general price levels of supermarket chains in their area ...

Price information is obtained from an independent monthly survey of approximately 500 products from 600 supermarket outlets across Australia.  The results of each monthly survey will be made available on the GROCERYchoice website on the first business day of the following month.[5]

... The basket prices for each supermarket retailer represent the average weekly basket price for that retailer within a specified region, and not the price for a specific individual supermarket outlet.[6]

2.8        Under the 'Meat & Seafood' and 'Fruit & Vegetables' basket categories, the website added the caveat that:

The quality [of some meat products / fresh fruit and vegetables] can vary between supermarket retailers over time.  Some differences in quality may exist and this should be considered when making price comparisons for this basket.[7]

The 61 regions

2.9        The GROCERYchoice website stated that:

... the [61] regions have been selected to ensure the survey is conducted across geographically identifiable areas, which are relevant to the lifestyles and shopping practices of Australian consumers.[8]

The opening page allowed a consumer to enter their postcode, or click on an interactive map of Australia to choose one of the 61 regions, in order to see the basket price results for their local area. For example, the way in which Queensland was divided into regions is shown below.[9] 

2.10      Clicking on the Brisbane region would allow the consumer to 'zoom in' on another map, divided into a number of other smaller, more densely populated regions.  The larger, non-metropolitan regions shown in the map above were counted as individual regions.  Clicking on the 'North Queensland' region (encompassing Cairns, Mount Isa and Cape York) would lead to a page showing the price of typical grocery baskets for a range of retailers across that region. 

2.11      In the same way, clicking on the region of Western Victoria—spanning Swan Hill, Avoca and Warrnambool—would lead to a page showing the following table[10]:

Basket	Coles / BI-LO	Woolworths / Safeway	Independents	ALDI
Meat & Seafood Basket	$26.43	$27.57	$27.73	n/a
Fruit & Vegetable Basket	$26.81	$24.77	$26.23	n/a
Dairy Basket	$15.79	$15.88	$15.87	n/a
Bread & Cereals Basket	$21.14	$21.44	$21.42	n/a
Drinks & Snacks Basket	$19.90	$21.07	$22.48	n/a
General Groceries Basket	$17.24	$17.99	$18.77	n/a
Household & Personal Care Basket	$26.81	$26.59	$28.14	n/a
Total of above baskets	$154.12	$155.31	$160.64	n/a
Basic Staples Basket	$79.51	$75.58	$78.99	$59.84

Preliminary scoping

2.12      As is the case with all budget bids, Treasury provided the Government with advice as to the merits or otherwise of the GROCERYchoice proposal.  However, this did not involve detailed economic modelling nor quantitative analysis prior to the allocation of funds for the website.[11] 

2.13      The ACCC undertook a scoping study, looking at different ways to set up the website, with 'obvious trade-offs in cost [and] reliability' and provided options to the Government as to the website's parameters (e.g. regional sample surveys were recommended in preference to individual census monitoring of every large supermarket[12]).  The website that was launched in August 2008 was 'fairly close' to what the ACCC had put to the government as being preferable.[13] 

2.14      In its February 2008 advice to the Minister on how best to establish a dedicated website to monitor and compare grocery prices, the ACCC provided details of its recommended model. The size and content of the baskets would be determined and weighted using data from the Australian Bureau of Statistics' Household Expenditure Survey as well as sales volume data to ensure that the dollar figures displayed on the website reflected approximate weekly expenditure by a typical household.  The ACCC proposed that:

... about 30 to 50 items would be selected for each sub-basket in each month, so that in any month about 300 to 400 of the 500 items are used in the basket, and 100 to 200 are not.  Over time, the contents of each basket is rotated through the entire list of 500 items, so that it changes gradually over time thereby assisting in maintaining basket confidentiality, but making sure the baskets can be sensibly compared over time.[14] 

2.15      The ACCC recommended that 500 items per store struck an appropriate balance between the need for a representative basket and the cost of data collection.[15] 

2.16      The advice to the Minister also showed that the ACCC had considered options which involved the monitoring and publishing of prices of around 50 individual products at individual stores (either a fixed list or a varied product list from week to week):

The principal advantage of both these options is that they would provide information on the prices of individual products on a store-by-store basis and in that sense would be highly transparent.

The principal disadvantage of such options would be that very large quantities of data and other information would need to be regularly collected and processed.[16]

2.17      The ACCC considered that the disadvantages (outlined below) substantially outweighed the advantages of these alternative options and ultimately recommended against individual product and store monitoring:

The monitoring system could be beholden to the continuing good will of the participating supermarkets, as it would only be effective with prompt, accurate and regular provision of data by the stores.

The burden of data provision would likely be significant for smaller supermarket chains. Coles and Woolworths would likely be able to comply readily with such a data request. However, such a request would likely be burdensome on individual IGA stores (which is essentially a decentralised franchise operation) and other independent supermarkets, and possibly unworkable for ALDI (which does not stock many of the items that would be considered standard in the other chains). The likely result would therefore be exclusion of IGA and ALDI (and other independents) from individual store monitoring. 

A further disadvantage of [the fixed list option would be creation of] incentives for chains and stores to manipulate the prices of those items, in attempts to be cheaper than competitors.[17]

2.18      The committee heard that prior to the 2007 election, the ACCC had sought advice from Informed Sources, a data collection agency, on the likely budget for a grocery price monitoring website:

They asked us to give them an estimate: if they were asked by government to put in place some sort of grocery monitoring, how many stores and what number of products did we think it might be across Australia?  We gave them an estimate of those costs, and it ranged between $2 million and $2.5 million.[18] 

Costs

2.19      Table 2.1 is taken from Budget Paper No. 2 of the 2008–09 Budget and shows that $12.86 million was appropriated for GROCERYchoice[19]:

Table 2.1 GROCERYchoice 2008–09 Budget appropriation

($m)	2007-08	2008-09	2009-10	2010-11	2011-12
	1.6	4.0	4.1	3.1	-

2.20      Treasury tabled a breakdown of total payments made to date in relation to the GROCERYchoice website ($7.7 million), which is at Appendix 3. Treasury estimated that the contingent liability associated with the termination of the contract with CHOICE was around $700 000 but advised that this amount had not yet been settled with CHOICE.[20] 

2.21      Around $73 000 has been spent on legal costs in relation to the GROCERYchoice website contract.  Treasury stated that legal advice for the period December 2008 to February 2009 had been mainly for the purpose of assistance in drafting the contract.  From June 2009, ongoing legal advice has been sought on a number of matters:

• advice on possible contract amendments which were not ultimately pursued;
• advice on the Government's right to terminate the GROCERYchoice website contract; and
• Treasury's rights and obligations following the termination and matters flowing from it.[21] 

2.22      The ACCC spent $3.6 million before responsibility for the website was transferred to Treasury, along with the remaining $9.2 million in funding.[22]
The breakdown of the $3.6 million expended by the ACCC was explained as follows:

$486,000 was in salaries; $64,000 was in salary oncosts; $60,000 was the accommodation and fit-out for the relevant staff involved; $1.4 million was for the data set-up and the data collection ... and $1.5 million was for the development and establishment of the website and the associated IT architecture.[23]

Collection of information for the website

2.23      In its advice to the Minister in February 2008, the ACCC stated that it had considered whether collection of price data should be done by survey or by requesting scanner data directly from retailers. Its recommendation was to undertake independent price surveys, noting that requests to supply data may be burdensome on smaller independent supermarkets. The ACCC also suggested that the use of scanner data could be seen (in the eyes of the public) to compromise the perceived independence of the monitoring process.[24]   

2.24      The ACCC was questioned about the collection of price information and the length of time it would remain on the GROCERYchoice website. The information would be collected over one month and then would be displayed on the website for one month.  When asked whether it was possible that the information on the website could be up two months old, the ACCC responded:

At the time that it goes on the website it would be about two weeks old ... You would have the survey completed, say, on day one and within two weeks of that the information would go on the website. It would remain on the website for a month, but in the meantime the next survey would be undertaken and that would overtake.[25]

2.25      Further discussion of the outsourcing of the data collection for GROCERYchoice appears later in this chapter.

The level of usage of the website

2.26      Treasury provided the following information on the number of website hits and page views recorded:

Table 2.2: Hits and page views on the GROCERYchoice website

Time period	Hits ('000)	Page views ('000)
August 2008	7,500	3,150
September 2008	811	446
October 2008	632	104
November 2008	844	111
December 2008	836	106
January 2009	293	105
February 2009	246	64
March 2009	204	52
April 2009	260	61

2.27      Treasury advised that page views are the best indicator of the usage of the website, as one page view is equal to one webpage, and a hit is equal to one graphic.   The number of hits is therefore greater than the number of pages launched and also varies depending on the number of graphics on the website.[26] 

Survey methodology

2.28      The 'Survey Methodology' webpage explained how the price data collected was used to create the prices for various baskets:

Each of the product prices in the basket are ... multiplied by an appropriate expenditure weight, reflecting the importance, in terms of average weekly expenditure, of that product compared to all products within the basket. These data are then aggregated by supermarket retailer and region to create the various basket prices.[27]

2.29      To ensure the quality and integrity of the data published, the website stated that the ACCC had engaged 'an independent product tester':

...to ensure that the grocery products in each of the baskets across supermarket retailers are of the same quality and therefore allow meaningful price comparisons to be made. This is necessary because the various baskets include a range of private label products that need to be matched to other brand or private label products available from other supermarket retailers.[28] 

2.30      Regarding the basket prices on the website, the ACCC was asked what mechanisms were in place to ensure that a grocery retailer could not engage in deliberate manipulation of prices to ensure that its basket of goods would be misleadingly presented best on the site. The ACCC responded:

To prevent the manipulation of the price survey and the published results the list of specific products and supermarkets included in the survey remain confidential to the ACCC. The sample of products and supermarkets included within the survey are also changed on a regular basis.

In addition to this, while there are approximately 500 products included in the price survey each month, a smaller sample of these products are used to calculate the basket prices published on the website. Changes to this sample are made so that the products contributing to the basket prices are not the same each month.[29] 

2.31      Referring to a caveat on the GROCERYchoice website (Note: the towns and suburbs included on the maps do not necessarily reflect where the GROCERYchoice survey is undertaken), Senator Bushby at Senate Estimates hearings raised the issue of different stores of the same supermarket chain pricing differently in various suburbs, asking:

Given the size of the regions – even the metropolitan regions that span a dozen or more major suburbs – how is a reasonable shopper meant to decipher the cheapest 'chain' in their area when the actual supermarket sampled may have been 10 stores away, or even hundreds of kilometres away?[30]

2.32      The ACCC acknowledged that:

Prices can vary between stores within the same group or chain, reflecting in particular the proximity to other local competitors.  For example, if a Coles store and a Woolworths store in the survey are in close proximity, the relative prices at both are likely to be affected. This is taken into consideration when selecting the supermarkets for the GROCERYchoice survey so that the average for the region is a representation of relative overall grocery price levels for each supermarket chain in that region.[31] 

2.33      (The practice of 'geographic price discrimination', where a different price is charged for the same product at different retail locations, is the subject of a bill currently before the Senate Economics Legislation Committee.  The Trade Practices Amendment (Guaranteed Lowest Prices – Blacktown Amendment) Bill 2009 would require major supermarket chains to charge the same prices at any two locations within 35 kilometres of each other.  If this bill were passed, or if other chains followed ALDI's lead and voluntarily adopted national pricing, it would make a GROCERYchoice website both much more useful and cheaper to compile.) 

2.34      The ACCC also stated that stores in regional and rural areas could be compared on the grounds on that they had:

...many similarities, including on the one hand the higher cost of transporting dry grocery products from central warehouses, and on the other the ability to locally source fresh produce.[32] 

Evidence of the website's effect on grocery prices

2.35      For the six months from July to December 2008, the ACCC's website published price data in 61 regions per month.  Of the 366 regional data collections, Woolworths was the cheapest in 198 regions, Coles was the cheapest in 160 regions and independent supermarkets were the cheapest in eight regions.[33]

2.36      The general consensus from retailers was that the website, while operating under the ACCC, did not have any discernible effect on grocery prices:

Senator BARNETT—Did it have any impact whatsoever in terms of putting downward pressure on grocery prices which was, I think it is fair to say, the government’s intent prior to the election? Presumably, this was a response to Labor’s promise prior to the election? Was there any evidence of downward pressure on grocery prices?

Mrs Osmond—Not to my knowledge. The thing that will affect prices is competition in the marketplace.[34]

2.37      At Senate Estimates, the ACCC responded to questions about the whether the website had any measurable effect on grocery prices while under its control: 

... over the three-month period of the site having had its surveys and the results of the surveys shown on the website, there has been a change in the relative positions of Coles and Woolworths ... In the first survey ... in 52 out of 61 of the regions in Australia, Coles was cheaper than Woolworths ...

In the last survey Woolworths was cheaper than Coles in 59 out of 61 sites. That suggests that there has been some movement on the part of Woolworths in its pricing practices, and we are aware ... that there were certain indications given at senior management level at Woolworths that they were not happy with being priced at a higher price than Coles on the GROCERYchoice site. We have also seen a closing of the gap between the independents and Coles and Woolworths in some of the pricing differences that have been present.[35]

2.38      This is suggestive of GROCERYchoice putting downward pressure on prices.  The ACCC expected it would do so, albeit to a modest extent:

To the extent that the GROCERYchoice website is intended to provide transparency to consumers and thus increase the competitive process, it would have a small influence. But you will recall the opening paragraphs of the grocery inquiry noted that the issue of competition was but a very small factor, like one-twentieth of the influence over grocery price increases over the past five years.[36]  

2.39      Professor Allan Fels, the former ACCC Chairman, and Professor David Cousins, recently commented that 'despite its severe limitations' the grocery price monitoring activity undertaken by the ACCC:

...seemed to have some beneficial impact for consumers, as the major suppliers seemed to respond to the favourable publicity for ALDI's low prices by also attempting to be the lowest price setters in a region.[37]  

Value of the website to consumers

2.40      At Senate Estimates hearings, the ACCC was asked whether the introduction of the website had 'tipped the balance of large chains versus consumers in favour of the latter.'  The ACCC responded:

Consumers now have an indication of which supermarket chain offers the cheapest prices across-the-board and not simply the lowest prices of 'specials.' This information was not previously provided by supermarket chains.[38]   

2.41      The ACCC also argued that despite the website not disclosing from which individual supermarkets the samples were taken, GROCERYchoice was still useful to consumers:

The survey is designed so that consumers do not necessarily need to know the individual supermarkets included in the survey in order to make use of the information provided. The basket prices presented on the website reflect the average price for a retailer within a specified region and not the price for an individual supermarket outlet. Over time the website provides a picture of which supermarket chain offers the cheapest prices in a particular region.[39]

2.42      The relevance of the website was questioned by Senator Joyce:

Senator JOYCE—As you are aware, Woolworths have said themselves that they have up to 4,000 price changes in a week. How relevant is your GROCERYchoice website when that is the fact—when it is at a snapshot in time and when one of these organisations asserts that they can have 4,000 price changes within their basket of goods in a week?

Mr Samuel—It depends on what is sought to be achieved by the website. If you look at the changes in relative pricing that I identified before between Coles, Woolworths, the independents and ALDI then the fact that there are changes each week in respect of so-called specials and other items would not be that relevant. What this does is provide a snapshot comparison on a monthly basis of pricing across 500 products. Those products are not known to the stores concerned. It does give an indication of their relative competitive pricing levels.[40]  

2.43      The ACCC also stated that despite the fact the data on the website could be up to four weeks old, 'the website provides a statistically significant indication of the comparison of the supermarkets as to who is more or less expensive'.[41]

2.44      A recent analysis of grocery prices at Coles and Woolworths stores in the Canberra region suggests that an 'average' figure for a chain is not necessarily of assistance to shoppers.  For example, while the two chains' average prices throughout the region were similar, a more detailed analysis revealed that one chain was cheaper in one part of Canberra, while the other was cheaper in another area.[42] 

2.45      Senator Xenophon suggested to the ACCC that the data on the website, given that it was only updated monthly, had the potential to be misleading to consumers:

Senator XENOPHON— ... for instance, in the first month, Woolworths could be the cheapest overall chain in a region and that led to consumers going to Woolworths in the next month. But what happens if in the next month Coles was the cheapest? It is a bit of a dud deal for consumers.

Mr Cassidy—That was one of the trade-offs in the design of the website and the amount of money to be spent on it. Clearly, the more regular the information, the less chance of a consumer going to one store when the price relativity has shifted.[43]

2.46      However, it could be argued that if the major supermarket chains did cut prices in response to the way they were represented on the GROCERYchoice website from month to month, there still may have been a good outcome for consumers. 

2.47      The ACCC also stated that the website was designed to provide information on grocery prices alone but noted that:

The website specifically mentions that price is only one of the factors considered by consumers when deciding on where to shop.  Consumers can take price information on the website into account with a range of other factors important to them when determining where to shop, including the location and accessibility of supermarkets to them, the quality of fresh produce, the product and produce range, and their assessment of the quality of service.[44]  

2.48      At Senate Estimates, the ACCC was asked whether it could guarantee that it was not favouring ALDI or the larger retailers through the GROCERYchoice website.  Mr Samuel told the committee:

I can guarantee that the survey process is a process that is bound by protocols of integrity and that there is no favouring or disfavouring of particular groups ...

As I said when the website was first launched, there are likely to be three reactions. One is to say, 'We're not faring well with this website and we will simply ignore it.' That probably does not go down too well with consumers.  The other reaction is to adjust the price to become more competitive and that appears ... to be what Woolworths has done over the past three months ... The third, and you would say almost the most obvious, course of action for those who are not faring well with the website is to denigrate it in every possible way. It has been pretty clear where the denigration has been coming from, but that is part of the process.[45] 

2.49      While accepting that the ACCC's 'mystery' basket approach was intended to prevent 'gaming' by retailers to achieve the perception of a cheaper store, the National Association of Retail Grocers of Australia (NARGA) argued that this also ensured that the information on the website was essentially irrelevant to consumers:

... no consumer was in a position to judge from the information on the website that a basket they might wish to purchase actually coincided with any of the items in the supposedly cheapest basket publicised.[46] 

2.50      The Retail Traders' Association of Western Australia asserted that the original GROCERYchoice website did not take into account consumers' shopping behaviour and displayed information that was effectively meaningless:

Understanding the consumer's habits would also have shown the complexity of the project. Consumers by nature do not travel extensive distances for food and grocery necessities and generally restrict their shopping for these items to within a 5 to 10 kilometre radius of their homes.

... To be useful to the consumer, the price information must be real-time, accurate and relevant to the shopping precincts they frequent. Averages, historical data and other non-specific pricing data are useless, even misleading and totally irrelevant.[47] 

2.51      The assessment of the ACCC website by Associate Professor Frank Zumbo, of the School of Business Law and Taxation at the University of New South Wales, was scathing:

...I have to say with considerable disappointment that GROCERYchoice website as operated by the ACCC was a complete waste of taxpayers’ money ... Fundamentally the website was flawed in its design. The information was too generalised. The thing we need to remember is that consumers shop locally. The GROCERYchoice website as operated by ACCC was broken into 61 regions which were very large regions ... It went from one side of the Sydney metropolitan area to the other side, some of the regions, and the reality is that consumers operate on a three to five kilometre radius ... Information was out of date; it was only collected once a month.[48]

Committee view

2.52      The ACCC's GROCERYchoice website was designed to fulfil a hollow election promise to put downward pressure on grocery prices.  However, it is clear that the aims of the website were never going to be achievable, with the sharply declining number of hits demonstrating that GROCERYchoice was of little or no use to consumers.  The poorly-designed ACCC website collected data by regions, some of them covering tens of thousands of square kilometres, bearing no resemblance to real-world consumer shopping patterns. 

Data collection - Informed Sources and Retail*Facts

2.53      The ACCC outsourced the data collection for GROCERYchoice to Retail*Facts, an arm of The Bailey Group Pty Ltd, which is a sales and marketing company.[49]

2.54      NARGA's submission pointed out that the ACCC's report of consultancy contracts let during the 2007–08 financial year to the value of $10 000 or more lists The Bailey Group as having conducted a 'Research Survey on Grocery Prices'. The cost of this 'restricted source' contract is listed as $5 135 650 (NARGA notes it was the largest consultancy for the year, about ten times that of the second biggest.)[50]  NARGA commented at an inquiry hearing that a more realistic figure for the data collection contract would have been 'well under $1 million.'[51]

2.55      A submission from Informed Sources, a data collection agency, also called into question the ACCC's decision to award the data collection contract to its commercial rival, Retail*Facts: 

The ACCC quickly and in our view correctly determined that there were two Australian companies who had the demonstrable experience, capability and management coverage to collect the data necessary to feed the GROCERYchoice website:

• Informed Sources – a company that, at the time, was in a heated debate with the ACCC, its Chairman and the Government over the potential introduction of FuelWatch. In addition to its collection capabilities, this company had extensive experience in the construction and running of a consumer pricing awareness portal (MotorMouth.com.au). Informed Sources bid $1.975M (excl GST) for collection of data for the GROCERY Choice website.

• Retail*Facts – a respected company with an extensive network of collection staff but (based on documents obtained under FOI) with the potential to fail the ACCC’s [request for quotation] needs for confidentiality/anonymity because of the deployment of that network and its likely overlap with its existing collection services for big Supermarket companies. Retail*Facts bid $4.669M (excl GST) for collection of data for the GROCERY Choice website.[52]

2.56      Informed Sources asserted that the ACCC spent 2.3 times the amount of money necessary on its data collection tender. 

2.57      The ACCC gave evidence that it had been under some time pressure to develop and launch the GROCERYchoice website:

Mr Cassidy—The government was keen for the website to be up and running as soon as possible ... We were working with an indicative time of having the first collection done so it could be released in early August.

... Senator BARNETT—So the government gave itself a self-imposed deadline to require it to be established within that six-week period. Did you advise them of the obvious cost differential? Did you advise the government of the implications of their push to rush this forward and to have it up and running so quickly?

Mr Wing—No. It was a policy and we had a budget so we just ran within that.[53]

2.58      The ACCC supplemented its response to the committee's question as to why it did not inform the Government of the cost difference, stating:

The procurement was conducted in accordance with Commonwealth Government procurement policies regarding value-for-money ...

The ACCC did not advise the Government of the difference in price between the two quotes because while they were both assessed to be within the budget for the program, only the Retail*Facts quote adequately provided for the delivery of services within the timeframe required.[54] 

2.59      Due to the timeframe set by the Government, the ACCC had some doubts as to whether Informed Sources had a data collection team that was 'ready to go'.  It was also noted that the ACCC had received several quotes and in the end had accepted the second cheapest tender:

We received a number of quotes ranging in price from about $2 million to well over $10 million. Our concern was very much that there was quite a major collection to be done—with approximately 600 supermarkets and approximately 500 products per supermarket—right across Australia and the first one would have to be up within six or seven weeks ... We looked very closely at the proposals by Informed Sources and others. At the end of the day, we were not convinced that there was a ready-to-go field force and we thought a fair amount of recruitment would have to be done. That was a great concern—that is, that there would be a great risk to the ability to deliver the data and high-quality data in time.[55]

2.60      Senator Barnett suggested that the ACCC based its decision on the Government's 'non-discretionary and non-flexible' deadline for implementation, to which the ACCC responded:

Yes. When we say ‘non-discretionary’, we were given a task and we did it, so it was not actually in our frame of reference or thinking, if you like, to sit there and think, ‘Well, perhaps we should push this back for six weeks or whatever.’ It is like a lot of times where agencies or departments are given a task by the government and they say this is what they have.[56]

2.61      Senator Pratt raised the possibility that the ACCC could not have had any foresight about which of the organisations tendering may or may not have been able to meet the prescribed timelines:

Senator PRATT—So it is quite usual in a tender process that you have to judge the tenderer according to the criteria in the tender? There are no second chances; if you do not make the standard then you fall away? You have to turn to your other tenderers to meet the tender, don’t you?

Mr Brocklehurst—Effectively that is the process in terms of the risk management decision you have to make: the quality of the tender, the submission, whether the times can be met, costs and so forth. It is all done as a risk management answer, effectively, in terms of who the preferred provider would be.[57]

2.62      Informed Sources' view is that the GROCERYchoice initiative was a 'failure of process' on a number of levels.  Mr Alan Cadd, Managing Director of Informed Sources, argued that excessive haste and the need to meet an 'illusory' deadline contributed to poor decision-making and design:

...there appears to have been a misinterpretation of the minister’s objective of having a GROCERYchoice website operating ‘as early as 1 August 2008’. That became ‘absolute implementation by 1 August’. In all of the freedom of information papers that we have been able to obtain, this simple policy interpretation has caused an unnecessary reduction in the time available, leading to rushed decisions and poor design. Surely, a reasonable approach would have been for a practical assessment of the time frame and for the departmental head to approach the minister with alternatives to reconfirm time frames against costs.[58]

2.63      He also referred to an 'inappropriate culture when handling taxpayers' money' at the ACCC, stating:

If there is one positive to come out of GROCERYchoice, it should be that every government employee in every department henceforth realises that the Westminster system is founded on a strong public service able to suggest and recommend to ministers best courses of action and not merely that they should spend the money if it is within budget. In our opinion, this was the ultimate failure of process.[59]

2.64      Responding to this criticism, the ACCC suggested there was a degree of hypocrisy in Informed Sources' willingness to tender for an initiative that the company believed was 'fundamentally flawed' and not worth pursuing as a public policy objective (see Informed Sources' comments on the effectiveness of GROCERYchoice in chapter 8):

I find it a little bit hypocritical for these people to come in and say that we should learn on what they regard as folly when they were more than happy to put in a bid and take the money for what they considered was folly. I find a certain—I would not want to say lack of integrity, but it is a bit hard to join the dots there when we are being accused of one thing and yet the company accusing us were going ahead more than happily in putting in a bid for what they think is a nonsense and a folly, and they were more than happy to take taxpayers’ money if we had chosen them.[60]

2.65      Informed Sources criticised the ACCC's tender process as 'the most rushed and frantic assessment process of something as significant as this that we have ever been involved in'.[61]  It appeared that the ACCC 'did not know what they were going to do, and that was reflected in the RFQ (request for quotation)'[62]:

All we had to do was provide the data to them. It was not in any way clear about what the baskets were and what the collection points were. We had an almost ridiculous circuitous argument with them. They asked, 'Where have you got staff?' And we asked, 'Where do you intend collecting this data?'  They said, 'We don't know yet.' I said, 'Let us see where you are going to collect this data, what supermarkets you are going to collect this data in, and then we can answer your questions about field force.' But that never transpired. That was right up until days before the decision was made.[63] 

2.66      Informed Sources stated that its bid for the contract, despite being the cheapest, had included a normal 40 per cent mark-up, and that based on the company's previous track record (including data collection for the introduction of the GST), there was no reason for the ACCC to doubt that Informed Sources was capable of delivering on the contract:

We were so confident with our ability and indeed with the prospect that Retail*Facts would not be chosen because of their Woolworths affiliation that we had no reason whatsoever to drop our regular margins.[64]

2.67      The apparent haste with which the ACCC designed and tendered for the GROCERYchoice data collection was also criticised by Associate Professor Zumbo:

That is just astounding ... Was the government ever told that their haste would cost taxpayers that additional $2.7 million? How long was the delay: was it a week, two weeks, three weeks? ... I just cannot believe that the haste can justify the huge additional expense ...[65]

Use of a separate field force

2.68      The ACCC's key concern in its assessment of the two bids ultimately hinged upon the issue of recruitment of a data collection field force. Informed Sources' submission drew upon the ACCC's tender evaluation documentation (obtained under the Freedom of Information Act 1982) which suggested that Retail*Facts won the contract primarily on the basis that it would be able to use its existing team of data collectors, whereas it was too risky to engage Informed Sources given that they were proposing to recruit an entirely new field force. Informed Sources argues that this decision effectively glossed over the crucial issues of confidentiality and anonymity:

The Informed Sources’ approach was to deploy a completely independent field staff team specifically employed and appropriately constrained with confidentiality provisions to ensure no unintended signalling of the survey to watchful retailers or suppliers. The winning tender’s approach was favoured (at least in part) by the ACCC panel because it made use of existing staff. Many of these Retail*Facts staff would have had long term relationships in the retail industry and indeed could have had dual or multiple working relationships and responsibilities. A merchandiser who worked in a supermarket for a supplier who now works for Retail*Facts simply can not turn off the existing friendships and working relationships built with the supermarket and the supplier over many years of service. “Chinese walls” in this case would be impossible to police and leakage of GROCERY Choice survey details highly probable.

Conversely, the Informed Sources’ approach would have seen a dedicated team whose job it was to avoid detection with no crossover of responsibilities.[66]

2.69      At the hearing, Informed Sources stated that, as well as dispatching a completely separate field force, further measures would have been put in place to ensure confidentiality:

We would have also clouded the products that they were collecting with non-collection products so that they would not have been able to determine which products were available. We also would have probably done some random collections at other times during the month, again, to try and confuse the issue. If these people were only being dispatched once a month then that would be too big a signal. All of those audit type provisions are just a natural way of doing business for Informed Sources.[67]

2.70      Asked to respond to Informed Sources' claims, the ACCC stated:

We made what we considered to be a well-informed judgment ... Even today I do not think anything Informed Sources have said would change our concern over whether they had the ability to actually get the workforce on the ground in the time they had suggested.[68]

2.71      Senator Xenophon questioned the ACCC on whether Informed Sources' previous track record was taken into account when assessing their GROCERYchoice tender. The ACCC replied:

Definitely. We would not have put as much work into assessing them, we would not have invited them back and we would not have asked the supplementary questions if we did not feel that on one level they were technically capable of doing the job. We went to those extra steps because of that. There is a notion that we did not go a bit further. We could have just looked at the initial proposal and said it is going to be a problem because they did not have the people on the ground. We could have gone straight to Retail*Facts. But we did not—and that was because of our relationship.[69]

2.72      However, the track record of Informed Sources was ultimately only one aspect of the risk judgment made by the ACCC in relation to its tender assessment:

When we get the tenders like that, we cannot sit there and just say, ‘Okay, we have all these doubts, so we’ll run with somebody’s track record.’ If in six weeks or two months time they were not able to provide that work, that would be end of my job. The thing is that you sit there and you have to make those judgments in terms of the information in front of you. You can quantify some risks. Most risks are more a qualitative judgment.[70]

2.73      The ACCC also assured the committee that Informed Sources' comments in relation to the GROCERYchoice data collection tender would in no way prejudice any future tender bids or arrangements with the ACCC.[71] 

Confidentiality and conflict of interest

2.74      The committee questioned the ACCC about Retail*Facts' simultaneous data collection contract with Woolworths, the safeguards in place to prevent any inappropriate disclosure of information, and the potential for conflicts of interest with such arrangements. Acknowledging that it would have been a serious concern if the information collected for the GROCERYchoice website was being leaked to Woolworths, the ACCC said that Retail*Facts had provided confidentiality undertakings for each of their data collectors:

It is not realistic to think, ‘If someone else is using somebody to do something then we will not.’ That is why we have safeguards and obligations in our contracts about confidentiality and so forth. It would almost be inevitable that some of the same data was being collected. Given we are collecting across 500 grocery items, there would be some commonality but only some in the sense that Woolworths or whoever else would be interested in some of the same items and some different items.[72]

2.75      The committee questioned Woolworths about the nature of the data collection activities contracted out to Retail*Facts:

...they provide a backup mechanism to our stores that do price checking with their local competition. So they have a range of stores and a range of different baskets that our data collection guys ask them to go out and do price checking on. They also do other work for us, like compliance checking to make sure that stores are putting the right tickets on the right products with the right specials on the right day of the week ...

...there are only a few companies in Australia that specialise in that sort of work, obviously, and there are only a few who do it very well. You want to get it right, because if they give you the wrong price and you set your price wrong then you are out in the marketplace and your customer misses out. I could only assume that they have in place the right sort of framework to be able to service multiple clients with the right confidentiality. We definitely demand it of them when they are doing work for us. If any supplier is doing work for us and a major competitor we have got all the right contractual arrangements in place, obviously.[73]

2.76      Woolworths also stated that it had not been aware that Retail*Facts had been doing work for the ACCC's GROCERYchoice website. Woolworths was not necessarily surprised, however, that Retail*Facts had also been engaged by the ACCC given that there are few companies in Australia that undertake data collection work.[74]  Mr Robert Hadler, General Manager of Corporate Affairs at Coles, said that in his view it was 'unusual' for the same data collectors to be doing price collecting for both Woolworths and the ACCC.[75] Mr Andrew Hall, Director of Corporate and Public Affairs at Woolworths, commented:

I do not know whether I would say it is unusual. Again, I would probably just put it in the context that the value of our contract with them is far less than the value of what I heard the ACCC was spending on them. Given the size of the price monitoring that I know they would have to have undertaken already for GROCERYchoice 1, one would assume that they were using a workforce far in excess of the people that they were using for us.[76]

2.77      Senator Xenophon questioned the ACCC further on the issue of confidentiality:

Senator XENOPHON—But would that contractual obligation also say, ‘We don’t want you collecting data for Woolworths at the same time’?

Mr Cassidy—No, we did not impose that sort of restriction.

Senator XENOPHON—Do you think you should have?

Mr Cassidy—No, I do not think that is commercial reality.

Senator XENOPHON—Did the ACCC review the Retail*Facts data collection processes—for instance, did the ACCC ever accompany Retail*Facts teams during data collection runs?

Mr Wing—No.

Senator XENOPHON—Do you think you should have?

Mr Wing—No, we contracted people to do this work.

Senator XENOPHON—I am not sure whether I am missing something here. You have the same team collecting data that could have conceivably collected data for both Woolworths and GROCERYchoice and you do not see any potential conflicts of interest there?

Mr Cassidy—Yes, and that is why we had confidentiality requirements in the contractual arrangement.

Senator XENOPHON—Is there any way that the ACCC could ever find out if there was any leakage of confidential information by Retail*Facts to Woolworths? How would you know?

Mr Cassidy—That is a good question. Basically, what we want to protect is what was in our basket. Over a period of time of running the website if we started to suspect that the prices on our baskets were being manipulated then we would start to wonder how the information as to exactly what is in our baskets got out. In the time we have been running the website, we had no indication of that and we were crossmatching the data looking for outriders and for data which did not seem to be consistent. We saw no evidence that there was any manipulation going on or that any information had leaked.[77]

2.78      The ACCC stated that the majority of collectors for GROCERYchoice were not used to collect price information for Woolworths:

There were a very small number of exceptions to this in remote regional areas. However, in these limited instances the price collections for the ACCC and Woolworths were undertaken in different weeks and were never performed in the same store.[78]

2.79      Price collectors had access to the ACCC's list of products for only two days.  They would download the product list using a personal digital assistant (PDA) network the night before the price collection task and access to the list would be removed following completion of data collection.[79] 

2.80      The ACCC's contract with Retail*Facts also specified that 'all internal control procedures, processes and practices would be open to ... an agreed audit test'.[80]  When asked whether any in-field checks had been carried out, the ACCC responded: 'we never got to that point because we did not see any evidence of anything being wrong with the data'.[81] 

2.81      The ACCC explained that it undertook a desktop analysis of the Retail*Facts data:

... we were crossmatching the data looking  for outriders and for data which did not seem to be consistent.  We saw no evidence that there was any manipulation going on or that any information had leaked.[82]

2.82      Associate Professor Zumbo commented:

The fact that they included an audit provision is good audit practice ... It is disappointing that they did not at any stage undertake that audit. You do not have to check all the data, just spot checks.[83]

2.83      He also raised the potential for Retail*Facts to have 'made a killing on this contract':

If you have the same data collection team collecting for Woolworths and Woolworths is paying you for that, and then you have the same data collection team collecting for GROCERYchoice, and the government is also paying you for that, you are killing two birds with one stone—you are getting paid double for doing the same thing.[84]

Retail*Facts' response

2.84      Giving evidence to the committee, Retail*Facts argued that its data collection offer to the ACCC was one of complete integrity and transparency.  Retail*Facts had disclosed its existing relationships with Woolworths and ALDI and had listed them as referees in the tender documentation.  Retail*Facts also highlighted the confidentiality undertakings signed by its collectors and explained the systems in place to ensure data confidentiality:

... we developed proprietary systems to provide specific product information to the price collectors and had that information available on a limited time. In most instances it was no more than 48 hours when that information was then taken away from their identification. Also to ensure confidentiality, specific and separately designed databases were used for each customer ... The notion that there was one database—that there was a price collector in a store collecting across a multiple number of our customers—is absolutely wrong. Every price collection that was undertaken was undertaken in a dedicated way. By way of explanation, every customer has different price audit requirements in relation to products, timings and the specific stores that are required. So there is no commonality apart from the fact that they are collecting price.[85]

2.85      Retail*Facts described its extensive internal data validation processes, including its own 'mystery' audits, undertaken monthly by a supervisory team on a minimum of 10 per cent of stores to review the data collected.[86]  Retail*Facts stated it was unaware of any in-field audits undertaken by the ACCC.[87] 

2.86      The inference that Retail*Facts could have undertaken 'dual collecting' for both the ACCC and Woolworths was staunchly rejected as 'ridiculous':

It not only shows a lack of understanding but a complete disregard for the way Retail*Facts does and will continue to do business, both with integrity and transparency.

A point that supports this is that throughout the 11 months that Retail*Facts conducted the price collections for the ACCC we visited some 6,255 stores across Australia and we collected details on approximately 2,400,000 products, which totalled, for the record, nearly 12 million at the points. It also should be noted that during the price collection process, throughout this massive undertaking, there was not one instance of dual collecting presented to us or to anybody that I am aware of. We are very proud of and would put on record the job we have done with the ACCC.[88]

2.87      When asked whether the same field force undertook the data collection for both the ACCC and Woolworths, Retail*Facts responded:

Mr Marshall—We did not use a completely separate field team. The point I would like to make, though, is that the way our business operates is that every activity that was undertaken for the ACCC, or for any price audit, is what we call a dedicated activity. That particular price audit person would be in the store and only operating for that particular customer on that particular day.

Senator BARNETT—Was it the same field force?

Mr Marshall—There would be instances where some people would be collecting on behalf of the ACCC and those same people, in a different situation—a different circumstance—and different stores, may be collecting on behalf of Woolworths. The point I would also like to make is that those persons do not have the knowledge of who they are collecting on behalf of.[89]

2.88      Retail*Facts also stated that the ACCC had queried 0.0001 per cent of their results (around 50 to 100 products) per check.[90] 

Potential for 'gaming' the basket

2.89      In its advice to the Minister in February 2008, the ACCC acknowledged the potential for problems when price surveyors are sent out into the field.  To alleviate risks, the ACCC proposed to:

... liaise closely with the price survey firm to ensure that price surveyors are thoroughly trained and that adequate and comprehensive contingency plans are in place ... The ACCC would need to ensure that data problems are dealt with appropriately (e.g. the goods are not on the shelf, or are a different size) and that surveyors maintain confidentiality and are not 'followed' by supermarket representatives (which is a common problems for this kind of work).[91]

2.90      CHOICE commented:

If it is true that Woolworths came out on top most of the time, that should spark questions with the ACCC about, perhaps, the basket of goods they were putting together or the methodologies that were being used ...

One of the problems with the ACCC site is that it only looked at a basket of goods of just under 45 items ... I suggest it was quite easy to game that basket. It is easier to work out what is in the basket of 45 items than it is a basket of 1,500. You could argue that, if you had a basket of 1,500 or 5,000 and supermarkets want to game those prices, that could be in the interests of consumers.[92] 

2.91      Regarding the secrecy of the ACCC's basket contents, NARGA commented:

It is my belief—I have no evidence of this—that people other than the tight circle would know what that basket was, if not in its entirety at least on some of the items. Very simply, as a retailer you just had to stand beside the person collecting the data and you knew that it was 750 gram-packets of Weet-Bix that they were taking the price of. You could not keep it secret forever.[93]

2.92      Associate Professor Zumbo also drew attention to the potential for gaming or manipulation of the basket of goods, while also pointing out the 'Catch 22' situation were the basket's contents to be rotated too often:

If it was rotated all the time, that would have provided some measure of feeling that perhaps there was some anonymity. But if at any point the rotation stopped and there was the same basket from month to month, it would not have taken long for the industry to know what was going on, particularly given that Retail*Facts also collected for Woolworths ... The other problem is: if you keep rotating the baskets, you are not comparing like to like from month to month, so there is a downside to rotating the basket too often. If you do not rotate it enough, they game the system; if you rotate it too much, consumers do not have a point of comparison. That is the fundamental flaw in the design of the ACCC website.[94] 

Committee view

2.93      The committee has serious concerns about the thoroughness of the ACCC's evaluation process for the GROCERYchoice data collection contract. The time pressure that the Government placed on the ACCC to launch the website clearly led to hasty decision-making and little consideration of the potential saving to the taxpayer of $2.7 million (the cost differential between the two data collection bids). 

2.94      It appears that at least $2.7 million could have been saved if the Government had been more flexible and kept its eye on the ball. The launch date for the website was arbitrary and politically motivated. 

2.95      While not suggesting a lack of integrity on the part of Retail*Facts, the company that won the contract, the committee is disappointed by the ACCC's apparent indifference to the risks inherent in Retail*Facts' simultaneous data collection activities for Woolworths. 

Recommendation 1

2.96      The committee recommends that the Commonwealth Auditor-General investigate the tender process undertaken by the Australian Competition and Consumer Commission in relation to the data collection contract for the GROCERYchoice website. 

2.97      The ACCC's decision not to undertake any in-field checks of Retail*Facts' price collection, as authorised by the contract, is particularly concerning to the committee. This lack of due diligence on the part of the ACCC leaves open the possibility that the integrity and secrecy of the GROCERYchoice data may have been compromised.

Recommendation 2

2.98      The committee recommends that the Australian Competition and Consumer Commission take more care in the future to monitor and assess the performance of contractors that undertake data collection on its behalf.

Navigation: Previous Page | Contents | Next Page