Google's Meridian is built upon the foundation of the previously released RBA/LMMM materials. The developments include geo experiments to ingest into the modeling, as well as detail on reach for YouTube. The emphasis on triangulation via A/B testing to enhance MMM accuracy is a strategy we are well-versed in ourselves and offers a good base to start from. However, it is crucial to note that while Meridian provides a step forward in measurement, it remains just a tool—a sophisticated one that requires expert hands to wield effectively. 

At Media.Monks, we pride ourselves on our robust internal platform that is industry-leading in terms of speed and functionality. Meridian gives a step up for brands who are just starting off in their MMM journey, helping them move away from last click to better quantify media uplifts.

From an experienced modeler’s perspective, these are some of the key points to consider with Meridian:

• The methodology behind Meridian is solid and makes sense around the emphasis on triangulation, which enhances the accuracy of the results.
• However, experienced econometricians will be essential for operating Meridian effectively in-house. Brands must ensure their teams possess the expertise to source the right data, build the models to reflect the real world, and translate data insights into actionable ROIs and response curves, or they risk making flawed decisions from the outputs.
• As with all MMM initiatives, data quality remains a critical factor in whether or not you’re adding value or making accurate decisions. Having accurate and full data across all drivers of sales (media, price, promotions, seasonality, climate, etc.) is critical for MMM. Strong data foundations also gives a significant advantage, whether brands are utilizing Meridian or any other technology.
• Effective communication within organizations is key to driving traction and implementation of MMM strategies, and explaining models clearly and effectively is key for any MMMs success.
• The launch of Meridian represents a shift away from outdated attribution models towards a more accurate, incremental media valuation approach. Even if it isn’t the best-fit tool for all brands, it is another step in the industry’s maturation, especially in the wake of cookie deprecation and changing privacy legislation.
• Smaller clients with simpler data structures, such as ecommerce clients spending less than $2 million USD on digital media, will benefit from this tool as an entry point to the world of MMM.
• Some clients may question running their media measurement on a platform from a media owner

In conclusion, Google's Meridian offers a solid starting point for less complex brands looking to enhance their measurement capabilities via a framework. Increasing the usage of MMM can only be good for the industry as a trusted tool to measure and optimize media. That being said, hard work is still needed in attracting econometric talent into the marketing world to maintain model accuracy and increase adoption of these methodologies. At the end of the day, a model is only as good as its modeler: you can have the best model in the world, but if it's not fed with accurate, high-quality data or delivered clearly to key stakeholders, it's not going to be trusted (and therefore, adopted) in an organization.

A good step forward, but still more to do on the talent front. See our post on apprenticeships to learn what we are doing to address this.

For more information on how we can help with your marketing effectiveness measurement or Market Mix Modelling, visit our Measurement page or contact us.

In the past, the idea of providing an automated MMM solution has terrified me. Blackbox solutions, which lack transparency, are as unsettling to me as they are to clients. However, the challenge lies in the fact that clients now demand speed and relevancy to ensure MMM thrives as it should. They require more models across their portfolio, greater granularity to account for channel fragmentation, and faster results and recommendations based on the latest campaigns and market conditions. If MMM cannot keep up with these demands, clients may resort to using easily accessible information, risking the possibility of making incorrect decisions based on inadequate data.

Therefore, I believe it is our role as experienced consultants to harness the power of artificial intelligence (AI) and machine learning (ML) to meet these needs, whilst simultaneously ensuring we still have a thorough understanding of the decisions which are being made within the process. We must guide the machine, set boundaries for AI, and sometimes intervene to provide information that the algorithms may not currently know or be able to find. AI and ML advancements should be utilized to build the core models, streamline hypothesis testing, and handle the heavy lifting. And we need to appreciate that a model itself provides limited insights. We (the consultants) must bridge the gap between the model and actionable recommendations, translating the statistics and the numbers into a language that clients understand and can implement.

The best approach combines the speed of machines with the detailed craftsmanship of the econometrician. It is a fusion of AI capabilities and the expertise of consultants that yields the most valuable outcomes. If you find yourself lacking in either aspect, I encourage you to get in touch with us. We would be delighted to discuss our approach further, ensuring that you benefit from the best of both worlds.

AI and ML offer immense potential for MMM, but they must be leveraged in a thoughtful and supervised manner. By guiding the machines, setting boundaries, and providing human expertise, we can unlock the full power of AI while ensuring accurate and actionable recommendations. Let us embrace the synergy between technology and human insights to drive success in the dynamic landscape of marketing.

For more information on how we can help with your Marketing Effectiveness measurement or Market Mix Modelling, visit our Measurement page or contact us.

Firstly, the importance of data in driving decision-making has become paramount. Businesses recognize the need for robust data-driven insights to navigate the complex marketing landscape. MMM provides a solution by quantifying the impact of various marketing activities, enabling businesses to make informed decisions based on solid evidence.

Secondly, the fragmentation of marketing channels has made decision-making increasingly challenging. With a multitude of platforms and channels available, businesses are seeking ways to measure the impact of their marketing investments accurately. MMM offers a holistic approach, allowing businesses to understand the effectiveness of each channel and optimize their investments accordingly.

Moreover, the rapidly changing economy poses unique challenges for businesses. Factors such as inflation, consumer confidence, political stability, global conflicts, and oil prices can greatly impact business forecasting. In this dynamic environment, clients are eager to utilize the latest intelligence to make intelligent decisions. MMM provides the means to analyze and adapt to these changing circumstances, enabling businesses to stay ahead of the curve.

Structurally, the attribution landscape has been undergoing significant changes. The deprecation of cookies, the rise of walled gardens, and the increasing digital investments have necessitated a more comprehensive and incremental approach to measurement. MMM has evolved to meet these demands, offering agility and granularity that align with the needs of today's market.

Gone are the days when MMM was a slow cruise liner, calmly sailing through the seas of marketing. It has transformed into an agile and adaptable tool, capable of navigating the challenges posed by channel fragmentation and rapid economic changes. MMM allows businesses to quantify what is working, accounting for the latest circumstances and driving data-informed decision-making.

As the phoenix rises from the ashes, MMM has risen to the occasion, demonstrating its resilience and ability to deliver meaningful insights. In a world where marketing strategies must constantly adapt, MMM stands as a powerful tool, guiding businesses towards success in an ever-changing landscape. Embracing MMM means embracing the future of marketing, where data and insights reign supreme, and agile decision-making holds the key to unlocking growth and profitability.

For more information on how we can help with your Marketing Effectiveness measurement or Market Mix Modelling,visit our Measurement page or contact us.

I am here to say that nobody puts MMM in a corner.

MMM not only quantifies the impact of various marketing activities on key performance indicators (KPIs), but also provides a measure of the effect on business performance of other controllable and external factors, such as price or distribution changes, competitors' actions, economic climate, and can even answer the question we all love to discuss, “How much of our performance is really down to the weather?” 

As MMM provides this panoramic view of all the key drivers, it is important to ensure it never operates in a silo if you want it to deliver its full potential. Here's why:

1. Cross-Functional Collaboration: MMM involves analyzing vast amounts of data from various teams. By demonstrating the benefits of MMM in terms of additional insights and recommendations, it encourages greater engagement from teams. This collaboration leads to a more comprehensive understanding of marketing effectiveness and drives better outcomes.

2. Influence Beyond Marketing: MMM has a significant role in shaping commercial decisions, particularly in pricing and promotions. Informed decision-making in these areas, such as identifying optimal price points, understanding price elasticity, and evaluating the impact of promotions on sales and revenue, empowers businesses to strike the right balance between profitability and customer demand.

3. Engaging Finance Teams: MMM often provides budget allocation recommendations across media channels, campaigns, departments, different brands in your portfolio as well as different markets. Involving finance teams ensures that recommendations are implemented and beneficial across the entire organisation. This collaboration quantifies the business decisions in both the short and long term.

In conclusion, MMM should be integrated across the whole business. By breaking down silos, fostering collaboration, and incorporating MMM into the decision-making processes, businesses gain more accurate insights, make better decisions, and achieve improved marketing performance and overall results.

Soon everyone will be holding MMM results and recommendations aloft above their head in the middle of the boardroom.

For more information on how we can help with your Marketing Effectiveness measurement or Market Mix Modelling,visit our Measurement page or contact us.

1. Will the measurement show the incremental uplift of media?

If the MMM models are measuring media but don't include the impact of other factors—such as Covid, seasonality, economic effects, etc.—then it is not providing you with an incremental measure and the media effects will be overstated. 

Always ask what factors other than media will be included in the model, and the sources of the data they use, to ensure your results are as accurate as possible.

2. What period of time does the model cover?

MMM needs at least two, preferably three, years of data to ensure it is deriving an accurate measurement of media and not conflating it with factors such as seasonality or other longer term impacts such as economic movements. If you are getting results with a lookback window of three months, then it's very unlikely to be MMM and therefore it will not be incremental measures you receive. 

Ask how much historical data the provider will require. 

3. What is the KPI that is being modeled?

Ask what the “dependent variable” will be. This is the KPI that is being modeled, and should be the metric on which your business success is judged. A sales metric—such as acquisitions, sales volume, revenue or similar—is ideal, as you can convert uplifts into revenue, then use margin to get to profit which enables you to assess true payback to the business bottom line.

If it is just web visits or digital conversions, alarm bells should be ringing! 

4. How are you dealing with interactive channel effects?

Any model needs to be reflective of how things work in the real world. For example, brand media can drive consumers to search for your products or services, which then drives up paid search. This needs to be accounted for correctly in the model specification, as well as any synergistic effects between channels and media’s ability to drive online and offline sales. If these are not accounted for, it’s probably not proper MMM. 

Ask how interactive media effects are taken into account.

5. How are you testing for causality, collinearity and significance?

These sound like complex terms, but they are not as scary as they seem!

Causality states the directionality of impact, i.e. which way something impacts something else. For example, does brand media drive consumers to search for a brand or does volume of searches impact brand media performance? There are certain econometric tests that can be done, which help determine this and validate your results. 

Ask for a list of all the possible data variables they would like to include in the model as well as the processes they will use to determine causality. 

Collinearity occurs when two factors move in a similar way and it becomes difficult to separate their impact, e.g. if TV and radio were planned with a constant weight over the same four weeks, an MMM model would struggle to determine the impact of each of these separately. Occurrences of collinearity can be tested for and should be flagged by the modeler.

Ask what kind of tests the modeler will use to determine collinearity.

Significance tells the modeler how important each of the factors are in the model. You need to be careful when there is low significance (usually on low spending media channels), as this is where the modeler cannot be confident in the result—which should be flagged to the client. 

Ask at what statistical level media is considered, and how the modeler will flag lower measures.

6. What is your verified forecasting error?

To establish a verified forecasting error, information about how the KPI has performed over a period of time is “held back” or not revealed to the modeler. The modeler needs to then use their analysis to “forecast” what they expect the KPI results to be. The forecast can then be compared to actual sales to verify the accuracy of the model.

The aim should be to have an error no greater than 8%, with a sensible range being between 2% and 8%. Non-incremental models (e.g. last click or attribution models) are poor at forecasting.

Ask if they have any validated forecasts from previous clients.

Hopefully this has helped to give a steer on what you need to be looking for. When in doubt, rely on these simple questions to get a sense of whether your partner is robust and reliable.

For more information on how we can help with your Marketing Effectiveness measurement or Market Mix Modelling, visit our Measurement page or contact us.

Understanding cohort data.

The use of customer cohort level data allows us to examine and predict the value of repeat purchases over time. A cohort level analysis typically seeks to answer a question along the lines of, “If I bring a new customer in with these characteristics, how much value will they bring to my business in terms of continued purchases?” Typical characteristics that may be factored into the analysis could include type of goods or services purchased, payment method, device type and time of year the purchase was made.

One such example can be shown in the chart below, where we see new customers that enter in Month 1 (initial purchase month) and the value that we gain from repeat purchase of that cohort of customers over time.

This allows us to apply an average multiplier to any future new customer revenue based on the characteristics provided—showing us that for instance, new customers driven at a particular time of year, or through a particular category or payment type, are more valuable than others. One such example is shown below, where regardless of category, driving credit customer revenue is much more valuable than customers paying upfront in full. We can also see that it pays off much more in the long term to drive new customer revenue through category A in Q1 and Q4, whereas we drive more long-term value from new customer revenue through category B in Q2.

Leveraging MMM to drive profit.

These insights are interesting themselves, but drive meaningful action when combined with media costs, effectiveness and profit margins to optimize a media laydown. Rather than simply spending on media when our traditional “busy periods” are and adapting to that, we can instead optimize our media laydown to take advantage of customer lifetime value instead. 

This allows us to not only run scenarios that seek to drive profit but also scenarios to maximize long-term customer value in the most cost effective way.

For more information on how we can help with your Marketing Effectiveness measurement or Market Mix Modelling, visit our Measurement page or contact us. 

But how do you ensure marketing effectiveness measurement is robust, defendable and clear? Here are some top tips for you to consider.

It has to start with objective alignment!

The first step is to be clear on your objectives: what is the campaign’s purpose? Is it to drive up knowledge of the brand? Or is it a reminder to purchase again? Are we trying to increase the reliance on online as its higher margin? By being clear about what you want the campaign to do, and to which audience, you are starting to define which metrics you should be measuring against, and therefore the definition of the KPI you should track against (for example, increasing awareness if your goal is to increase brand knowledge). This article by data guru Avinash Kaushik is worth a read on further defining the right KPIs.

Set realistic targets!

Once you’ve chosen the KPI, you will need to set realistic expectations of how you expect the campaign to move the KPI. For this, look to the past to see how much the metric has moved; if there hasn't been much variation, then perhaps you shouldn’t expect a huge increase. If there is a lot of movement, what looks realistic in terms of uplift? Make sure you consider seasonality. For example, look at the three-year pattern: are there certain times of the year that are always up? If so, take that into account.

How will you measure?

With targets set, you will need to think about how you are going to measure the campaign before you deploy it. Some techniques include:

• Random control tests (RCTs) and geo testing
• A/B tests
• Measuring against a historic baseline
• Multi Touch Attribution (MTA)
• Market Mix Modelling (MMM)

Knowing which technique you will use helps you define the shape of the campaign. Knowing which technique you will use helps you define the shape of the campaign. For example, if you undertake geo testing, you will need to identify the most appropriate geographical area for your activity to occur within and a comparable area to use as a control. Meanwhile, for MMM, you will need to ensure you have sufficient media spend levels and variation to enable you to get a read on the impact.

Make sure the test spend has the following key attributes:

• Is there enough spend to move your KPI?
• Is it shaped so you can get as clear a read as possible (i.e. bunch it up, don't spread it out)?
• Is it at a time that will conflate with other impacts?

Execute and measure.

Once you have the right objectives and metrics, know which measurement method you are going to use, and the campaign is successfully deployed, it is time to then execute and measure. When it comes to measurement, be conscious of limitations of your measurement technique.

RCTs, geo and A/B approaches are easy to deploy, simple to understand and can be deployed internally. However, there are some limitations to these techniques, which can prevent them from giving a full read on the effectiveness of your activity.

First, there's difficulty getting a read on the "carryover" of the campaign (often called the memory effect), which is the effect the campaign continues to have after the campaign has finished. These approaches also struggle to measure the impact of specific media activity onto other channels; for example, running social activity can boost PPC. These findings are key when trying to understand a full view on your media performance.

These methods are also unable to provide information on scaling the results. A test spend of £20k in one region will not have the same ROI as a £5M national campaign. Be aware of diminishing returns. You can get around this by upping the investment in increments and continuing to test.

MTA is great at giving detail, can be relatively easy to set up, and gives you a good relative read. It is not, however, an incremental analysis, so it is not reliable for ROI calculations.

MMM is incremental and includes a full read on all drivers of your KPI. However, it is blunt (you need to spend at least £100K on a campaign), does not give as granular detail as other techniques, and can be expensive.

Considerations for in-housing measurement.

So, as a client what could you do yourself?

• RCTs, geo and A/B tests: Most of the time, there’s no real need for external partners.
• Multi Touch Attribution: Give it a crack, it’s fairly straightforward and you can use techniques like Markov chains. But you can only use traditional MTA for a short while: with third-party cookies going away, there is a longer term need to invest in cookieless solutions.
• MMM: Great to in-house if you have the scale, but you need to keep a team busy and fulfilled. This, therefore, only works if you are either an enterprise or global company spending over £100m on media.

Be aware of a couple watch-outs for in-housing.

• Don’t use data scientists for MMM—use econometricians. We’ve seen time and time again that where data scientists have been tasked with building MMM models, it very very rarely works. Data scientists and econometricians see things in different dimensions.
• Econometricians are hard to find, and harder to keep hold of. You need to make sure the work is varied and interesting for them to stick around.
• Make sure there is enough work for at least four people. Otherwise, if you are reliant on one econometrician and they leave, it's very hard to get someone else in to pick things up.
• Career progression: A consultancy can always get bigger, but an in-house enterprise team hits a limit. So then you have no progression and people leave. Or you move them into non-econometric roles but then you still have the problem with recruiting in a niche field.
• Method stagnation: There is less opportunity from your econometricians to learn new techniques from larger teams working on other clients. So there is a danger that your capability starts to fall behind the curve—unless you have really high staff turnover, which you’ll need to have a big enough team to support.

However, if you are large enough and have the right team, then in-housing can save a lot of future money on consultants, whilst keeping complete control of your data and models. To keep it moving, and to make sure you don't stagnate, consider using an external partner to help refine your MMM approach.

Summary

In conclusion, getting the start nailed down is critical in marketing evaluation. Align objectives with KPIs and how you lay down the campaign. Think carefully about which technique you will use to measure, and whether you do that internally or with external help. In these uncertain economic times, there has never been a greater need to get this right. Bon chance!

For more information on how we can help measure drivers of growth for your business visit our Measurement page or contact us.