A Brand’s Marketing Measurement Journey
Marketing measurement is not an activity rather it is a journey that many stakeholders start together to achieve certain business outcomes. Marketing measurement is called many names depending on what kind of business you are in and what kind of measurement you want to perform.
- Impact of marketing on Sales
- Benchmark return on investment ROI
- Measure the impact of campaign
- Optimize marketing mix
- Identify the right marketing budget
Many organizations measure marketing for most profitable or most marketed brands as traditionally it has been quite expensive to perform marketing mix modeling (MMM*) for brands that sold online as well as in stores. That led to infrequent and ad hoc measurement for many brands where brands measure every 18 or 24 months. Most of these marketing measurement projects are “pure reporting” projects and provide zero insights into future actions as these projects take a long time to collect data, run models and product hundreds of pages of power points for executive’s consumption.
Why should a brand adopt frequent marketing measurements?
When a brand moves to more frequent (half yearly, quarterly, monthly) measurement of marketing performance compared to ad hoc tracking; here are the direct benefits you witness:
- Develops discipline of looking at marketing activities
- Provides a consistent version of truth in what is driving business performance.
- Builds an institutional knowledge of key metrics
- Improves predictability of sales forecasting models
There are few indirect benefits that comes from the frequent measurements:
- Avoid the media waste while being more experimental & agile.
- Save time for middle managers who spend most of the time preparing reports for leadership.
Develops discipline of looking at marketing activities
By instituting frequent marketing measurements, a brand can create an internal process of ongoing reviews on market activities between business units and agencies.
- It forces focus on trending forecastable indicators of activity if an activity is going to be included in the model.
- It creates internal conversation and alignment about what metrics are the best indicators of tactics and/or activity. For example, on digital — is it spend, impressions, CTR — what should we be tracking and when?
Provides a consistent version of truth for business performance drivers
With frequent measurement, a brand is able to review YoY and QoQ drivers, that provides the team a capability to assess changes in performance over time.
While you are measuring frequently, make sure to choose a vendor that provides rapid measurements with granularity (no compromise on quality due to speed). With that, a brand will be able to look into media, distribution and promotions executions at such a deeper level. Ability to double click on all of the marketing activities at a campaign level also provides a clarity on what is working or not.
A brand is then able to find actionable insights that helps their media agencies improve media plans. For ex. over time, if you witness that YoY increases in media performance for a brand is driven by digital channels such as Display, Video and social; this insight will help the brand focus on high performing media channels.
Builds an institutional knowledge of key metrics
Over time, the team learns to look at ROI with respect to other metrics such as sales, spend and effectiveness to understand changes in performance of each marketing channel. Concepts such as ad-stock/half-life, flighting, decay should be taken into account during execution.
Based on the principles of readability provided by your marketing measurement vendor, the team should be able to execute experiments. Make sure your measurement vendor has the guard rails i.e. the data points, in the model, that help articulate confidence while measuring smaller experiments.
Improves predictability of sales forecasting models
A brand can align the quarterly measurement timelines such that outcomes (likely sales impact from marketing activities) can be fed into the internal sales forecasting process. This helps brands estimate the monthly sales impact for the next 2 years.
Continuous measurement helps brands make faster, better, data-driven decisions. By performing frequent (quarterly or half yearly, if not more) performance measurements, a brand is able to improve the overall predictability of sales forecasting models.
Key Learnings for a team building a measurement capability
There is just so much confusion when it comes to measuring the marketing. Many times stakeholders put too much focus on how much data they have (1st party, 2nd party, and 3rd party), how many tools are connected, what modeling techniques should we employ. All of that is really important, after all data is at the center but people, process, teams involved in this journey are the key to make this journey an overall success. Here are few tips:
- Start small & Set up the cross-functional team
- Don’t try to incorporate everything in the beginning
- Do not underestimate internal change management
- Treat output as dashboard vs. deck
- Choose a right measurement vendor
Start small & Set up the cross-functional team
Identify a brand that has sales and marketing channel investment that is tracked for at least one year.
Set up the team. It is very important to bring a project manager from day one. It is important to bring Analytics & Business leadership for the initial set of meetings. Here are some key stakeholders are their respective contributions that would make this whole engagement successful:
Business & Analytics Leadership plays a key role in vetting the model outcomes from a business point of view. Key aspects where leadership can add value:
- Role of marketing channels; expected ROI
- What can or cannot be done in terms of recommendations provided as an outcome of measurement exercise?
Analytics leadership has an important role to play in establishing the models early on and then monitoring the tweaks as and when they happen. Key aspects where leadership can add value:
- Keeps everyone in check when it comes to making model assumptions, validity of models (including independent variables, dependent variables), limitations to regression analysis (w.r.to to domain) and quality of predictions.
- Triage the issues around models & outcomes with vendors and the business leadership.
Project Manager plays the key role in the engagement by being the glue to all stakeholders in the process. Key aspects where leadership can add value:
- Ensure that data from various sources are provided to the measurement vendor on time.
- Facilitate validation of final data with relevant stakeholders.
- Consolidate questions/hypotheses that stakeholders may have.
- Get an alignment with business & analytics leadership on outcomes before sharing it with the larger audience (including agencies & brand team).
- Work with your measurement vendor to educate teams on how to use the outcomes for better decision making.
Bringing agency (ies) on board is the key as they would have to share all the media buying, engagement and outcome data.
Initially, a large amount of time will be spent in:
- Finalizing the media data from the agency side.
- Explaining the marketing measurement methodology (mostly done by measurement vendor),
- Understanding KPIs & how various channels are executed by the Brand & measurement vendor.
The relationship between agencies & measurement vendors would evolve from untrusting to supporting to enabling, over a period of time.
Data Platforms/Providers are sitting outside your company but would be part of this measurement process when it comes to data collection and underlying limitations and assumptions w.r.to data provided.
Once you have an understanding of which marketing channel, platform and/or tactics works or not, you would adjust the media spend based on recommendations from your measurement vendor.
In this process, you might get push back from various data platforms such as print, FB, Google through your media agencies. The measurement vendor along with your media agency might have to get on calls to explain the methodology and reasons behind the recommendations.
Make sure the measurement vendor is capable of assessing the media mix model specifically for the publisher to ensure that the right metric is used to model.
Sometimes data platforms might have to provide additional data (that may not be available to the agency) with conversion metrics or additional metrics that can support model thinking.
Don’t try to incorporate everything in the beginning
If you are executing “marketing mix modeling” engagement first time, focus on understanding the model outcomes on a high level first and then get into details one at a time. For ex.
- First spend time on understanding how “Display” marketing channel is working on a high-level and its interaction with other channels.
- Once understood, get to the publisher level.
- And then, at a tactic or creative level.
Trying to consume all the detail at first would be hard to digest and would distract the team from core learnings.
Based on the marketing execution, identify which of the activities should be measured as part of the activity. Sometimes you can measure outcomes of a particular activity within a week of execution while it might take longer for other kinds of activities. For ex. the outcome for certain promotion coupons/samples can only be read after about 15 months as opposed to the outcomes for digital execution where the impact can be seen in as less as a month.
While starting the measurement activity, understand the measurability guidelines provided by the measurement vendor. Be ready to exclude the marketing activities that do not necessarily comply with guidelines. need to be done for a certain period of time and at certain thresholds to allow for measurability.
Be ready for multiple iterations for the initial analysis, therefore, plan a longer time period.
Do not underestimate internal change management
Start with a process and then incrementally improve it by putting checks & balances based on the stakeholders involved & learning from the previous analysis. Always remember that it would never be perfect and you might have to focus on fixing things one at a time.
- Stakeholders might not comply with the timelines to provide hypothesis & business questions
- Data templates are changed without communicating to everyone involved in the process.
- Expectation to incorporate new data sets without advance notice.
- Few participants never open the document before meeting and few might need a week’s notice; balance out the need.
Once the deliverables are finalized, identify which downstream processes can take the outcomes from marketing measurement exercise on an ongoing basis to improve the overall business. For ex.
- Response curves, for all media and promotions channels, can be used in the forecasting process to integrate short-term performance of marketing channels in the last quarter.
- Share your outcomes and key metrics with various teams managing downstream processes to understand what kind of additional data can be brought into modelling to bring additional value over a period of time. For ex.,
- Unstructured data i.e. comments, reviews, digital conversations
- Measuring brand equity in addition to sales
Plan for training on “consumption of outcomes”. Focus on highlighting what is directional vs absolute information. It takes a few rounds of analysis to truly understand the outcomes.
Stabilize the process
Once there is an understanding of data sources & its owners, standardize the data collection process including the templates for extracted reports from various tools within the organization.
A lot of times, internal churn affects the data extraction process. Therefore, it is critical to document the extraction process to keep the process independent of people who own the data extraction.
Once the initial round of analysis is complete, templates for all sales and marketing channels should be shared to ensure that data is received in the same format for future analysis. As changing the data templates would affect the overall data vetting process. It might take additional effort for the modeling vendor to consume the data as many times the data management process is customized for each brand/client.
Finalize the process steps and deliverable formats. Keep discussing with all the stakeholders what information a deliverables helps them vs confuses them. This would help you incrementally improve the quality of deliverables while keeping the formats consistent over a period of time and across brands.
Any changes in data/deliverable templates as well as the process should be discussed and communicated upstream & downstream.
Tools & Templates
Build an inventory of tools that are the key to provide data for the measurement exercise. For ex., Data tracker — it is an inventory of all the data sources, data type/metrics, frequency and the owners of each data source.
Keep the process documented to extract the data/reports to avoid any attrition related impact on the process.
Educate stakeholders about the data templates and why it is important to follow the templates.
Once the outcome templates are finalized; work with the vendor to build the quick guide/cheat sheets to train the stakeholder on how to consume outcomes.
As an outcome of the measurement exercise, identify standardized metrics that would guide your team in making decisions:
- Attributed EQ volume/Sales: How much sales is driven by each channel
- ROI: How well is this channel performing? It should be looked at in conjunction with the spend and Attributed EQ volume and not by itself.
- Conversion KPIs: There are some non-sales metrics such as CTRs, Viewability etc. that are available in the everyday data that are correlated to sales/ROI. These KPIs can be used as indicators of performance.
- Effectiveness: Understand the performance of the channel with respect to the execution and not just spending. Especially when detailed spending is not available, this helps accurately understand performance
Treat output as dashboard vs. detailed power points
Leverage model outcomes in dashboards (browser based or excel based) rather than creating detailed PPTs. This would expedite the “time to action” by almost 2 weeks.
- Dashboards make it much easier to identify data gaps, anomalies, patterns and trends.
- Dashboards can provide a significant amount of detail for various time periods and KPIs by channel, sub-channel and any detailed execution requested for by the teams or agency (performance by coupon, performance by ad length etc.)
- Analysts/Brand teams/Agencies that are involved in the execution and decision making can get access to performance metrics and underlying data via these templates. It brings transparency about data at execution level
- Generate an executive summary (PPT) for business leadership interested in seeing answers of specific hypotheses.
Choose a right measurement vendor
When you are taking on a “marketing measurement” engagement the first time; here are some pointers to keep in mind while choosing the right measurement vendor
- Transparency: Would you have visibility into the modelling box or would it be a black box to you?
- Reflecting business landscape: Is the technology equipped to learn from business changes (adding new variables, dropping old ones, merging variables, baseline related changes, etc.) and deliver on-time; this might mean changing the models multiple times.
- Collaboration: Can the vendor work with your various agencies, coach them and support them in building the best media plans possible for your brand?
- Pushing boundaries: Capability to ask questions to business, marketing/media and analytics team to break group think.
Having measured almost $750Mn+ marketing spend globally in a typical year, I believe what brands truly look for in a measurement vendor is combination of speed, price and innovation.
Do share your thoughts on how you are taking your brand through the marketing measurement journey. Drop me a note at LinkedIn to discuss any of these points further.