How much does a demand signal repository cost?

Firstly, let’s establish what we mean by a Demand Signal Repository (DSR) – then we can consider different flavours, what the cost drivers are and suggest typical costs for varying degrees of sophistication and capability. Starting with some definitions:

  1. A DSR is a centralised database suitable for data about demand (very generic);
  2. A DSR provides a central location to store, manage and retrieve multiple different types of ‘demand signal’ – information that directly or indirectly describes demand for products or services; or,
  3. In the grocery retail / consumer goods space, a DSR provides a central, consolidated store of many different demand signals – acquired from retail customers, market data aggregators, social networks, internal systems etc. – and often provides a means to match these with corresponding supply signals to enable analysis of product demand and the ways in which it is affected by supply, price, brand sentiment, weather, retail channels and many other explanatory variables.

This article will focus on the third of these – DSRs designed for fast-moving consumer goods (FMCG) and consumer packaged goods (CPG) companies – which is where our expertise and experience lie. We will get to budgetary costs by the end but need to sketch out the landscape first.

What flavours of demand signal repository should I consider?


How small can we start?

At its simplest, a DSR could be nothing more than a single table of sales data collected from a single retail customer. For a startup brand producing a few products and selling these through a single supermarket chain, this will be sufficient as a start point. This could be managed in an off-the-shelf spreadsheet tool, like Microsoft Excel.

What should a mid-market supplier expect from a DSR?

For any business or any size, with multiple products and customers, it’s more likely that you will want to collect, manage and analyse much more customer data; not just sales figures, but stock positions, availability metrics, promotional responses, stock movements through your customers’ supply chain locations (stores, depots, distribution centres).

Many UK supermarkets, for example, provide suppliers with access to large datasets that describe daily trading performance in considerable detail. To handle multiple data files, from multiple customers, it’s likely that you will need to consider:

  • How to collect data automatically (especially if you are collecting data daily)?
  • How to store collected data in a reliable and accessible form?
  • How to collate and compare data across multiple customers (from multiple systems)?
  • How to translate customer-specific terms e.g. retailer products codes, week numbers etc?
  • How to access stored data in your preferred reporting and analytics tool(s)?
  • How to integrate DSR content into other business systems?


Modern DSRs will be able to meet all of these needs: collecting data from key customers, translating and aligning customer data into a consistent format, and presenting a unified data set to your analytical tools and business systems.

What’s the state-of-the-art?

The very best DSRs today do more than just collect, transform and store data for analysis; they enrich and complement the source data so that it’s increasingly valuable and useful in a wide variety of downstream business processes. A few examples:

  • Creating synthetic metrics for data sources that don’t include the ‘normal’ full set (e.g. availability)
  • Providing a decomposition of achieved sales to offer an explanation of underlying ‘base’ demand, long-term growth (or decline), seasonal and special events, stock shortage impact, promotional uplift (and cannibalisation)
  • Producing forecast and ‘backcast’ datasets that complement the original source data with counterfactual or explanatory data, useful for predictive analytics, forecasting engines and data science projects
  • Incorporating machine learning and artificial intelligence(s) to discover, highlight and automate best practices

You should consider what ‘value-added’ DSR features you require today (if any) and whether the DSR that you select is capable today or includes these as part of its roadmap.

What are the cost drivers of a demand signal repository?

Most modern DSRs – once beyond the simplest of home-grown databases – are cloud-based; that is, they are delivered as fully-managed applications on the internet. Gone are the days of requiring database designers, administrators and dedicated servers. The primary drivers of costs, for these cloud-based DSRs, are:

  1. Technology
    1. Data volume, variety and frequency
      Whilst storing data is relatively cheap (around $10 per terabyte, at the time of writing), collecting and processing large volumes of data can require considerable computing power. More data, collected and processed more often, will increase the compute costs for your DSR provider.
    2. Data enrichment and pre-aggregation
      Synthetic metrics, decompositions and forecasts are usually calculated using very large training data sets which, in turn, require large amounts of compute – albeit typically refreshed on an infrequent basis. The more sophisticated the enrichment, forecasting and pre-aggregation, the greater the computational cost.
    3. Integration points
      For some DSRs, integration points increase underlying cost-drivers, and so providers of such DSRs will typically charge for each integration point, e.g. your different business systems and analytics tools.
    4. Retrieving data
      Every time you run a report, an analytical dashboard or an ML model against your  DSR you are making demands on the DSR’s underlying data platform. This results in compute (query processing) costs which will be passed on to you in some form.
  2. Service
    1. Service implementation
      If you select an off-the-shelf DSR you will either deploy through your own team (requiring training, learning curve allowance etc.), through your DSR provider’s team or through an approved partner. DSR implementation effort varies wildly; depending on the nature of your deployment you can be up-and-running within hours (using pre-build connectors for major customers), or you may require weeks of planning, deployment and testing before production deployment. The most modern SaaS (Software-as-a-Service) DSRs tend towards rapid, low-cost deployment but do so at the cost of ultimate flexibility – they do what they do and no more. 
    2. Support
      All DSR providers will provide helpdesk services, to ensure that you receive support when required. Some provided graded helpdesk options – for example, enterprise-grade support on a 24/7 basis with a 4-hour SLA – for an enhanced fee.
    3. Training
    4. Platform maintenance
      All cloud-based DSRs, including those delivered under a full SaaS model, will incur costs to maintain their infrastructure, security protocols, software stacks, etc. They will either pass these costs on to you directly—through a maintenance programme—or incorporate them into their standard fees.
  3. People
    1. Some DSR providers also supply end-user tools for account managers, demand planners, customer service representatives and senior management. In these cases, the number of people using the DSR – and its associated tools – tends to affect both the Technology and Service costs that the DSR provider incurs. In most cases, more users equals more cost, so many DSR providers will license their products on a per-user basis or in tiers (e.g 5, 10, 25, 100 users etc

So, as with much in life, you tend to get what you pay for. If you collect, manage and query a relatively small number of demand signals from a handful of sources and use standard services (deployment, support and training), you might be looking at a few hundred pounds a month for your DSR, potentially including best-practice reporting and analytics tools. This can be a great starting point for innovative food and beverage startups, right through to scale-up challenger brands, and selecting the right DSR provider who will help you to scale your DSR as you grow is often the best option.

If you require an enterprise-grade DSR, collecting tens or hundreds of data files from customers across multiple markets and regions, and harvesting millions of demand signals daily, you are likely to want a range of tailored services and are probably looking at spending several hundred thousand pounds each year on your DSR and associated professional services.

What should I budget for a demand signal repository solution?

Due to the variety of DSR approaches, and the wide range of services that DSR providers offer, it’s difficult to provide a simple price – as we hope this guide has demonstrated. It is, however, reasonable to outline typical budgetary ranges for DSR services (all expressed as thousands of pounds £GBP):

ApproachDesign/planDeployMaintain (pa)Operate (pcm)
Home-grown, minimal5-1010-205-100.1-1
Home-grown, typical20-4040-10010-501-3
Home-grown, enterprise100-200200-60050-2005-25
SaaS, minimal0-50-5–0.25-1
SaaS, typical5-200-10–2-10
SaaS, enterprise10-5020-100–5-50


In summary, SaaS options tend to provide a faster and lower-cost entry point, with all of the technology already in place, pre-configured and maintenance included in the (monthly or annual) service fee. If you choose to implement your own design you will incur more upfront costs for design, development and deployment but, usually, with lower operating costs – although system and data transformation maintenance can prove risky and, in some cases, very costly.