
7a1d2a5a238ade8bc6b3d76473de82de.ppt
- Количество слайдов: 14
Using historic data to improve planning and forecasting TFM&A 2014 David Lockwood: Direct Wines Terry Hogan: Golden Orb Campaign Planning – Direct Wines
Context • UK planning based on multiple linked spreadsheets – Other countries followed different approaches – Budget process took up a lot of resource • No single reporting tool or data warehouse – UK business on different systems from the rest of the world – Assembling data for marketing meetings very time-consuming • Forecasting continuity sales cumbersome and inflexible • Different terminology/KPIs in use in different parts of the business – In some cases the same term being used to mean something different in different countries • Decided to build a global marketing system to support marketing across all countries Campaign Planning – Direct Wines
Outline – system fundamentals A single system to store and report on past campaigns – Automatic feeds from UK and international transactional systems One system for reporting and planning future campaigns – Learnings from past campaigns feed into planning future ones Shared database rather than Excel – Single version of the truth – One set of definitions, terminology – Standard, automated methodology for calculating/forecasting Campaign Planning – Direct Wines
Modules Sales reporting Forecasting Campaign planning Campaign Planning – Direct Wines
Campaign-level sales reporting Gross/net Orders – by day, order channel, response code. . . Comparison with budget/ plan Detailed campaign financials Numbers split out in various ways – Response code – Test – List/publication Products sold Campaign Planning – Direct Wines
Time-based crosstab reporting Aggregates the results of multiple campaigns over time Can split out results across several dimensions Typical ‘dimensions’ for reporting – Campaign – Response code attributes (activity, media type, list. . . ) – Time – various levels – Order channel Ability to view top-level numbers and then to drill into the detail Campaign Planning – Direct Wines
Sales reporting - considerations Need a feed of order information from transactional systems – Orders/revenue – Product costs – Aggregate numbers by e. g. response code, day – Doesn’t need details of individual orders or customers Probably a daily feed needed for campaign sales (overnight) – Other data can be updated/stored weekly Don’t over-complicate or request too much detail – There always trade-offs between detail and performance – Different levels of detail are appropriate for different measures Campaign Planning – Direct Wines
Campaign planning: summary of the process Volume X Response rate % Cases per order Orders Cases P&P, AOV and product cost Revenue /margin Existing customer % Response curve & channel split Recruits Phasing Application of direct/indirect fulfilment and marketing costs gives net contribution To phase the sales forecast over time, you need to build a sales profile/response curve Campaign Planning – Direct Wines
Building a response curve Choose a number of representative campaigns Express each response code’s sales in terms of days after start date Calculate the overall percentage of orders received by day 1, 2, 3. . . Build different curves for different campaign types and media Requires good data, correct campaign details Campaign Planning – Direct Wines
Phasing the campaign forecast Apply sales profile to the top-level campaign numbers – Plus order channel split if relevant – For Direct Wines, we generate a daily sales forecast by order channel The system can build an aggregate curve if the start dates are different Unless you have a very simple product range, this sort of tool is not appropriate for detailed product planning – Primarily a tool for planning marketing activity Campaign Planning – Direct Wines
Reforecasting a campaign after it has started The response curve can be applied to the actual sales to date to create a revised forecast of final sales Can apply a different percentage to orders through each order channel Ideally needs to be done every day for each response code and order channel Day-of-week adjustments may be needed for an accurate reforecast Changes in forecast flow through to customer service, merchandising, finance Campaign Planning – Direct Wines
Modelling non-campaign sales Direct Wines have a continuity business in addition to ‘standard’ customer marketing – The continuity business generates sales without an associated ‘campaign’ that needs planning – Sales to existing continuity members can be modelled on the basis of their existing memberships – We can start to forecast continuity back-end sales when planning recruitment or ‘upgrade’ campaigns The main continuity forecast is rebuilt weekly – However changes to recruitment or upgrade campaigns have an immediate knock-on effect Campaign Planning – Direct Wines
Continuity modelling A complex area – the full details are beyond the scope of this talk Continuity behaviour is quite predictable over time for a reasonably large group of customers Simple approach – apply a curve to the initial recruits Two enhancements: – Actions at cycle n depend on actions at cycle n-1 – For current cycle, we can use the actions received to date to adjust our estimate Campaign Planning – Direct Wines
Thank You www. golden-orb. ltd. uk @Golden. Orb. Ltd Campaign Planning – Direct Wines