BASES Analyst Challenge Client Solutions

Analyst Challenge Introduction

Welcome

BASES is Nielsen Innovation Practice. Go through the module to learn more about your role & responsibilities in Nielsen Innovation.

Analyst Challenge Learning Objectives


The Analyst Challenge is an opportunity to reflect upon challenging BASES clients´ questions and recommended answers. This training will help prepare you for difficult situations with your clients.

This training has been compiled in a matrix format. For each topic there are 1-2 matrixes of client questions with links to reveal more details. Please ensure you scroll down to see all of the questions and solutions.

BASES Client Questions

Study Methodology

Understanding the Client Business Problem is the first step of Study Design. A BASES study is not as simple as ordering a burger & fries, it is imperative to balance creativity with realism – to ensure we have the answers we need while recognizing some questions cannot be answered in a BASES study.

  Question 1  Question 2 Question 3
  The BASES sample doesn't match the population. How can this be accurate? 
Alternate client question: Why are BASES' Panel demographics (age, income) different than census demographics?
Is an Overquota of Prescription Brand X Users required to forecast sales of the OTC version after a switch? My product is a niche product, why do I have to test among a General Population (GP) when few of these consumers are category buyers?
Ask: Can the client provide target/non-target marketing plans?
Answer BASES sample represents consumers most likely to purchase this type of product (e.g. Principal Grocery Shoppers). Results among this group are a better predictor of sales than a representative sample based on the census. 

Socioeconomic details: If faced with questions about under-representation of the bottom social class, discuss purchasing power. While class E makes up 50% of the population, they represent <5-10% of sales. Matching census to volume forecast would decrease accuracy, therefore it is more important to interview those with purchasing power.

Yes, since prior prescription users are more loyal it is recommended to include a readable base of current Rx users. However, depending on the feasibility of interviewing that group (i.e., low incidence brand), category Rx users are interviewed as a proxy. It is necessary to interview all probable purchase influencers. 
More: In an Rx to OTC switch we typically test the concept among physicians and pharmacists (in some countries) in addition to consumers. Why?

  • Physicians developed product opinions and might continue recommending although not prescribing and some patients/consumers are reluctant to buy without first consulting with their physician. Therefore, this measures physician driven volume to switch and physician brand support
  • Pharmacists can also assess recommending intention
  • NOTE: Since the doctor is not on-hand to educate, consumers may not understand the difference between the Rx version and OTC. Different recommendations are likely (Example: Prilosec Rx recommendation for GERD and can be taken long-term, but OTC is only for heartburn up to14 days)
For many CPG products, often a third or more of volume comes from outside the target because products can attract category non-buyers. Marketing events will inevitably attract households both in and outside of your target. For Example: You could be testing super premium dandruff control shampoo that will attract regular shampoo buyers.
Consulting Tips Answer should include ‘we can boost if you are interested.’  - Positioning can be helpful. Position as a target sample with heavy diagnostic analysis among this group and boost of GP interviews for forecasting purposes ‚Äč
Applicability - Pharma only  -

 

  Question 4 Question 5 Question 6
  BASES doesn't do top two box placements anymore, while that's ok for food it's a problem with health care products where most of your sample will never suffer the ailment.
Question detail: Why did BASES switch top three placements? We need less concept interviews to hit our placements and fewer interviews means lower study costs for our clients.
You won’t quote a confidence interval on Pre-BASES, although when I test multiple concepts in the same marketing plan, concept sales estimates are within 30% of each other. If the difference is so close, isn't Pre-BASES useless to rank concepts? How did you get Source of Volume Analysis (SOVA) numbers? Does BASES use the chip allocation process?
Note to analyst: Under the chip game, up to 7 brands are chosen from aided and unaided questions to the screener.  Brands are chosen based on most used and never used.   Respondents are asked to allocate 11 chips between all pairs of brands (up to 21 pairs!) based on how much they like each brand (preference).
Answer Before expanding our placement methodology to also include respondents from the middle box who claim “might or might not buy”, BASES conducted extensive research to determine the appropriate data calibrations in database comparisons and the forecasting model.
More: Where we made changes in placements, our validation record - across all categories - has remained stable, suggesting it is possible to accurately interpret scores and forecast sales using this methodology.
30% is a real/meaningful difference.
We are confident in Pre-BASES estimates. Pre-BASES uses the same forecast system as BI/BII, therefore if the marketing plan provided is complete resulting estimates would be comparable to BI or BII. The marketing plan is also used to rank concepts.
Confidence interval responses depend on client or situation, but can include:
- We don’t quote a confidence interval because we use condensed marketing plans. 
- Pre-BASES uses a smaller sample size (and which may have been sequential monadic).
Considerations: 
  • With consistently similar results, consider innovating/ differentiating/ ideating products more.  A MBA could help identify white space opportunities/ areas of strength
  • Consider all variables in the Strategy, Consumer Viability, Financial Potential (SVP) not just ‘P’.  Concepts may play a different role (on strategy, defensive launch with parent brand heritage, strengthen the brand, stretch into a specific area) or may be more challenging to implement (one has high salience while another doesn’t).  This provides a ranking of the concepts across all measures aiding to make an informed decision on products to prioritise.
Prior to 1995, BASES used the chip allocation for SOVA but we achieved better results changing methods to align with Nielsen Sourcerer evaluations post-launch. BASES’ consumer sourcing claims are based on responses to the question, “If you went to the store to buy new (insert product name) and no varieties were available, what would you buy instead?” BASES II uses a share based approach in the after-use interview (asking the SOVA question multiple times).
Chip allocation isn't the best method for SOVA but, it is a valuable in the MBA, observing historical purchases. The goals and benefit versus cost are different. 
Consulting Tips - The client may ask you guess the confidence interval. Do NOT make one up (like +/- 30%). Assure the client you have the same confidence as other BASES tests (assuming the client provided detailed marketing plans and a finished concept) because we are using the same model.  BASES utilizes CHIP in both MBA and the Competitive SnapShot developed for UL. 
Applicability Does not apply to Pharma and countries that place Top 2 Box (Lat Am, some AP, some EMEA) - -

Interpreting Data

  Question 7  Question 8  Question 9 Question 10
  Purchase Intent is above average for my category but BASES predicted my new product would be a failure. Which results should I believe? I tested many concepts using SnapShot. It was clear that the Concept Potential Score (CPS) ranking was based almost entirely on frequency. This doesn’t help me screen and prioritize my concepts. SnapShot results said we were ready, the BASES II says we are a probable failure. Should I believe the SnapShot results? __% of consumers say they definitely or probably will buy at concept. How can you tell me that these scores are bad?
Answer Purchase Intent and Factors for Success measure two different things. Purchase Intent is an input in a BASES forecast and measures the financial potential of an initiative. Factors for Success help identify potential risks to success in market (consumer adoption). Success is more than financial potential. Success is the consumer adoption and financial potential
of an initiative, and the costs and risks of achieving it. These costs and risks need to be weighed against the potential of the initiative.
Consumer adoption and financial potential are usually consistent but sometimes they tell two parts of the same story. A weak result for either consumer adoption or financial potential may reveal a risk in market. When they don't line up, we can draw meaningful conclusions. Often, this seeming disconnect indicates the product will be highly cannibalistic or perhaps, only appeal to a niche group of consumers.
Important Note: No difference in purchase intent is a learning in itself!
First, remind clients they should be more reliant on the Success outcomes vs. CPS. Both consumer adoption and volume potential are important and need to be examined in the decision making process, but consumer adoption (Factors for Success) provides more insight on what needs improvement and how to improve it. Conversely, CPS doesn’t measure what/why. Therefore, comparing Success outcomes is a better way to evaluate concepts. If concepts have similar Success outcomes, then layer CPS scores to evaluate.

CPS (CPS is based on weighted purchase intent (all boxes of PI), units, price, and frequency...not just frequency. Frequency is important to volume (consumer involvement with the category, repeat potential), although other aspects are important to success, therefore it is important to prioritize what constitutes success. Empirically there are reasons to see the benefit from frequency, BASES determined CPS projects better when including frequency (double jeopardy, another vote for PI).
Discuss strategy: If CPS is similar across concepts, other criteria should be considered to help choose “winners” – consider using uniqueness, a loyalty measure,or the client’s cost (R&D, capital investment, etc) or profitability to make one concept more attractive.

The goal for a SnapShot test is typically to prioritize early concepts based on their overall potential. The goal with a BASES 1 or 2 test is typically to optimize one of the "winning" concepts from SnapShot and estimate year 1 sales. We include only the Factors appropriate to measure at each stage of development.
For example, early concept screening is typically done before the selling message is finalized. Therefore, Factors which measure the effectiveness of the communication are generally not included (though they can be added to any study). It is possible the additional Factors we measure in later stages of testing will reveal risks which were not appropriate to measure at earlier stages. Often times, poor product performance is the culprit for this scenario. Remember, strong performance on one Factor does not compensate for poor performance on another and an initiative is only as strong as it's weakest link.
PI is a volume metric and should be compared to the client’s goals. We should evaluate the concept as “above average” vs. “below average” compared to the database. (Combined with marketing, we evaluate the volume potential of the initiative which provides additional decision criteria to Factors for Success.)
More: Consumer overstatement is why BASES believes in using databases to provide perspective.  Although consumers overstate, they do so in a consistent and predicable manner.  Thus, with robust databases, by category and by country, we can provide context for data interpretation.   For example, in the US, the average top box purchase intent at concept is 20%, so a score of 30% is above average.  Conversely, the average top box score in the Philippines is 60%, so a 30% would be sub-par.  Norms will vary by category, country and methodology.
Consulting Tips - If testing across categories, you may see higher CPS for a more frequently moving category. Take the CPS into context when consulting with the client.  - -
Applicability - - - -


 

  Question 11 Question 12 Question 13 Question 14
  Is it fair to expect __% of people to say that a new product is “Better than Expected”?
Factors for Success: Focus on the Product Delivery factor.  What would it take to move the concept from failure/risky to ready or outstanding.
Consideration: Focus on the Product Delivery factor.  What would it take to move the concept from failure/risky to ready or outstanding.
Why don’t we have Factors for Success for Restagers? I’ve heard that the BASES Probability of Success is driven by the weakest link factor, but I have a failing factor and my overall score is risky. The client wants to know more about a Factor result. Should I show the data for the question(s) behind the Factor? For example, on Distinct Proposition, should I show mean Alternatives score and the stacked bar (% of responses in each box) result?
Answer Performance vs. expectations compared to the competition is the primary metric for predicting product performance in market. It is not enough to meet expectations. Successful products exceed expectations. By comparing the scores for your initiative to our database of like products, then comparing it’s rank to the observed thresholds for success we can predict the level of risk associated with product performance for this product.
More: Keep in mind, the % Better than Expected required to be market ready will vary by category, country and methodology.

Restagers have built-in product benchmarks: the comparison to the current product’s performance.  Unlike new brand and line extensions, there isn’t a recipe for success on a re-launch.
A brand considering restaging will typically have a couple of benchmarks where growth is critical, other areas to maintain performance/perception and expect no decreases on all other areas.
More: Re-launches can be volume focused, strategy focused, or both. For instance: 

  • Executing a strategic shift in perceptions vs the marketplace (e.g. move the brand to a premium tier) while maintaining or growing volume
  • Increasing volume by attracting non-buyers and maintaining existing buyer base (expansion strategy)
  • Changing product formulation maintaining all aspects of the franchise (product performance, volume, etc)

The BASES Overall Probability of Success calculation begins with the weakest link, but is most often higher than the lowest probability of the 12 Factors. Expect BASES Probability of Success to be slightly increased from your weakest Factor. Overall Probability calculation:

  • 84% weakest link 

  • + slight upward calibration 

  • + slight increment for blues

    More: The “Weakest link” analogy provides the desired consulting framework from a conceptual standpoint, although the factors cannot “compensate” for one another.  If you correct one factor and others are low, the overall probability will be comparable.

Show  the ‘What’ and the ‘Why’ slides.  The ‘speedometer’ slide presents the ‘what’ i.e. what is my story, how am I performing on this factor and how to get to ‘ready’ if underperforming or leverage a strength.  By presenting slides containing the measures for the factor, for example, the stacked bar on CALT and the dial slide on distinct proposition, you are effectively showing the ‘what’ twice – the stacked bar slide of a measure adds nothing and should not be included in your reports.
Example: Supporting slides should explain ‘why’ the factor performs the way it does.
For Distinct Proposition, demonstrate how the initiative is differentiated.  Does the initiative offer benefit-driven innovation?  (Use the report content to explain result, not to prove it)  Does it enable new usage occasions? 

Consulting Tips - - - Alternate Nielsen Solutions: FactorsQual (for UL) and Factors Advisor for more in-depth learning and actions. 
Applicability Does not apply to Pharma.  Does not apply to Pharma.  - -

 

BASES Factors for Success

BASES Factors for Success reports results in probability of overall success. This provides insight of the overall Probability of Success for new products based on real market cases. 

  Question 15 Question 16 Question 17 Question 18
  Do Factors for Success apply globally? What about differences by category and geography…shouldn’t the breaks be different?
Alternate Client Question: What are the Factors thresholds for my category/country?
How many countries and categories were the Factors for Success analysis cases based on? What is the breakdown by continent?
Note to analyst: Don’t get hung up on details. These lose sight of the big picture, we have the ability to provide accurate advice regardless of the situation (country, category, etc).
New Factors for Success: In overstatement countries how can we feel certain of results when the response scale is condensed at the top end of the scale? 
Note to the Analyst: The answer is essentially how we accurately forecast in overstatement markets.
How predictive is the Factors model for snacks? Of course no one "needs" potato chips, why do you ask that for indulgent categories?
Answer

Success breaks (Outstanding/ Ready/ Risky/ Failure) are the same for every concept we test regardless of category, country, and other classifications. Example: if a concept in the US needs to meet the 45 %ile to be ready on Need/Desire, so does a concept in GB.
Universal success breaks are possible because the foundation of success models are database ranks, ranking on a relative basis, equivalizes category or country differences and provides a universal prediction of success. What it takes to be successful (in terms of absolute scores) may differ between countries/categories however, what’s important is performance vs competition ( database ranks)
More: Although, our R&D indicated the most predictive success models use database rank.

  • Actual scores: absolute performance
  • Scores indexed to category medians
  • Database ranks: rank vs. competitors in BASES Key Measures Database (relative performance)
  • A rank + score hybrid

Submodel details: Sub-models were compared for each success model we built: continent, category, developing vs. developed, etc. Sub-models didn’t show any significant differences from the main model. Market factors were also included in the regression inputs, but did not emerge significant in any model.

In addition to the 600 launch cases, BASES conducted additional R&D on our full testing experience of 20,000 concepts with introSCAPE data. We feel good about the dataset. The profile is consistent with BASES overall global testing mix, which allows us to accurately advice regardless of the situation (country, category, etc). The dataset is well represented across:

  • Geography. All geographies represented, more than 30 countries spanning all continents
  • Categories. Proportion of cases in each category consistent with BASES overall global testing mix
  • Clients. Over 60 clients included
  • Study types. Line extension versus New Brand mix is consistent with our testing history

Our country/methodology-specific survey data conversions remain one of our key assets. We have the ability to account for overstatement, both large and small. Database ranks are predictive in all markets including high overstatement markets. We generate an accurate db rank resulting in a reliable success prediction. 
Factors R&D:

  • We’ve been able to mitigate overstatement with to identify measures that render a greater data differentiation; this was imperative in determining which (of the 78 potential) measures are used in the Factors for Success model.
  • Factors R&D investigated regional & category specific success models (as well as many other types of models) and none of them varied from the larger model in any significant way. This is the basis for the conclusion that the Factors apply to all geographies and categories.

More: For Factors for Success: Most of the measures are not as susceptible to overstatement, as measures like Purchase Intent, because they are behavioral in nature.

The Factors model "hit rate" is ~75% for all categories/countries. There is no meaningful difference in the overall predictability of the Factors model by country, region or category. 
We ask the Factors questions in the same order and same way on every study for two reasons. First, in order to database the results, the question must be asked in the same way every time. Secondly, the databased results are an input into our Factors model. We evaluated hundreds of questions in the Factors R&D and this series of questions (need/want) question asked exactly this way added to the predictability of the model. Remember, we use the maximum of the need & want questions in the Need/Desire calculation. This question was purposefully designed to account for category specific variations. So, for indulgent categories, the maximum responses may mostly fall to the "want" question and the model/predictability will not be impacted. 

 

Target Groups

  Question 19
  What mean score do I need among my target group to overcome a risky or Failure among the General Population?
Answer There is no such "score." The Factors R&D includes only the GP and "databased subgroups" such as parents & teens. We do not have R&D which links results among client-defined target groups to in market success. Any suggestion that this is the case would be misleading to our clients. We can provide relative perspective by comparing a client-defined subgroup to other studies which include that group but there are some watch outs. The definition of the group should be the same in order to compare results across studies. Also, even if the definition of a subgroup is exactly the same from study to study, the make-up of that group may differ. For example, brand A buyers in one study may be comprised of older consumers than brand A buyers in another study. In this case, different results between the tests may be caused by different levels of overstatement among different age groups, and therefore they are not "real" differences.
Consulting Tips -
Applicability -

Forecasting General

Volume is not the same as Success 

  Question 20 Question 21 Question 22
  What is your awareness? Aided, non-aided, self-identified? Can you tell me what you are forecasting my awareness to be?
Note to Analyst: Typically, we do not report our awareness estimate, as awareness weights are proprietary (we may provide a range for Year 1 tracking if pressed). Although, BASES previously provided awareness levels, so if the client previously worked with BASES (years ago but not recently) they may expect it.
Why can’t I compare Nielsen panel components directly to your forecast? BASES Probability of Success is Failure or Risky but the volume estimate says we will exceed the volume goals for this product. Which should I believe?
Answer BASES awareness estimates are most similar to aided consumer tracking awareness levels AMONG THE GENERAL POPULATION. Since most awareness tracking is executed among target groups (category buyers, etc.), use caution when comparing to BASES GP awareness levels.

BASES volume components do compare to Nielsen’s panel components after accounting for timing and coverage.

  • BASES Year 1 begins when the product reaches 10% ACV/distribution. Align Nielsen T&R and BASES Year 1 start dates. If ACV builds slowly, this can impact Year 1
  • Reconcile differences between panel, store and shipment data. Work with the client to determine in-market sales (e.g. Nielsen RMS scanned sales projections and client shipment data minus pipeline). Then, coverage adjust Nielsen panel to match sales.
Achieving volume goals is only part of success. The Factors for Success framework identifies risks to consumer adoption which volumetric questions do no measure. For example, this product may achieve its volume goals but it may be highly cannibalistic of the parent brand since it doesn't offer something additional or better than current brand's offerings. Users will just pick up this product instead of their usual brand purchase.
Consulting Tips Never react with non-flexible approaches ‘its proprietary’ or ‘we can’t.’ While this is true, a consultative voice should focus on why they might not need it (i.e. not useful to them, not matching what is tracked in market) and what we CAN offer (i.e. MPA perspective on their ad plan.) If pushed then resort to the proprietary nature.  Be prepared for the question, “so, how do you adjust?” Align with Nielsen counterparts on adjustments.  -
Applicability - Does not apply to Pharma or non-panel countries.  -
  Question 23 Question 24
  You're wrong on Year 2 learning. I've been the President of major CPG companies for twenty years and new products always grow in Year 2. BASES only knows how to model traditional mass advertising - you don't know how to forecast micro-marketing events like internet, out of home, and PR.
Answer BASES considers the new products as only SKUs that launched the first year, while many CPG companies launch line/variety extensions in Year 2. Line extensions are often accompanied by incremental distribution and added consumer interest, resulting in new brand volume increases. The initial items often lose marketing support in the second year, experiencing volume decreases.
More: Even without line/variety extensions, new products can increase in volume in their second year by attracting new triers (Increased marketing spending to Year 1 or added distribution) and retaining strong purchasers (high buyer repeat). Products with long purchase cycles and Holder/Refill initiatives tend to grow in Year 2.
BASES has been actively conducting ongoing awareness research.
 This includes validations of over 200 items with non-traditional media. A review of validations indicates our accuracy is consistent with or without internet advertising. Modelling non-traditional advertising is similar to traditional activity (i.e. GRP, TRP impressions).
We also have ongoing access to learning coming from other Nielsen divisions, Nielsen Media, BuzzMetrics, etc. In mid 2013, BASES Digital Task Force leveraged rich global Marketing Mix Modeling (MMM) data, improving digital media modelling (by updating best practices and creating a tool to incorporate digital inputs in BASES forecasting).
This is not a simple issue, but an evolving science; we will work with you to learn more.
ongoing awareness research: 
By 2009, BASES collected data on 160+ products in six countries. This, combined with Nielsen Media and BuzzMetrics data, reviewed claimed sources of awareness and source trends. Findings include:
  • Shelf presence always matters, especially for launches with small marketing budgets; even with typical media support, in-store is the single biggest source of awareness for new brands and line extensions
  • While declining as a source of awareness, TV advertising may be the best/only way to drive very high awareness levels, given its reach. TV’s decline has little to do with DVR usage (so far)—too few commercials are time-shifted and skipped to matter
  • Internet continues to grow in importance. CPG manufacturers spend less on Internet than other industries (which makes sense), but also less than market potential might suggest
  • Outdoor seems to be an emerging resource, partly due to new, eye-catching forms (like skyscraper ads.)
Consulting Tips - -
Applicability Does not apply to Pharma.
Regional Note: In Mexico and Lat Am, generally distribution isn't maximized until Year 2, offering continued growth in Y2. 
-

Forecasting Components

  Question 25 Question 26 Question 27
  My current product has a 40% repeat rate – why does my test brand have a 28%? Frequency is an important part of the model...we've looked at the tables, and there is a lot of variance. Do you account for possible outliers in the data set? You mentioned concept scores are used to estimate trial.  You've also said trial is the key driver of sales.  How can concept results have a weak relationship with long-term sustainability? 
Answer Repeat tends to be lower in the launch year due to 2 factors.
1) Trial often occurs later in the first year while distribution and awareness build. Therefore late triers often repeat after the product's first year in market.
2) The more a customer has purchased an item, the more likely they are to buy again. Established brands tend to have a lot of buyers making their tenth, eleventh, and twelfth purchase, while new products have many buyers making their first or second purchase during Year 1.
Example: Process specific details depend on the situation. For example:
  • For a new item forecast wee use in-market measures within the category and country to determine relevant benchmarks (using double-jeopardy patterns of penetration vs. purchase frequency within a category)
  • For a restage forecast, we use the current form as a benchmark (relating Restage vs. current claims to current in-market measures)
  • For screening multiple concepts, we smooth certain claimed values (considered unstable) to an average to limit the effect of instability
Frequency is an important consumer measure in our volume forecasting. However, it is also the most overstated measure. Adjusting for overstatement controls differences that might be caused by outliers. BASES has access to consumer data from Nielsen partner companies to identify if forecasted measures appears to be (without good reason) atypical. 

Although concept purchase interest is not correlated with long-term viability, purchase intent at concept can indicate potential size of the business. Manufacturers can “buy trial” – that is, get a consumer to try a product once…but, the secret to sustainability is retaining as many triers as possible. 
Factors for Success amendment: 

  • Based on the Success R&D, many factors at the pre-use phase are important to overall success. Ten of the twelve Factors for Success (9 of 11 for Unilever) are at the pre-use phase, and failure on any of these can result in short-term failure in market
  • Long-term sustainability is driven by product performance. Based on the Success R&D, product uniqueness and value are the most important to long-term sustainability. Although the pre-requisite is Product Delivery, initial product expectation fulfill concept promises and consumers' expectations


 

Consulting Tips - When interpreting Key Measures we classify measured values into quintiles, which can be quoted to give additional perspective.  -
Applicability - - -
  Question 28 Question 29 Question 30
  Since we didn’t place product in this study, how did you estimate a repeat rate for our product? In a presentation a senior BASES person stated free standing inserts generate awareness. How did you handle the FSIs in estimates? How much do they contribute to awareness? Does your model account for shelf dynamics? Our Ipsos (Novaction) representative indicated that your model doesn't take the shelf into account.
Answer When we do not have product placement, we base repeat projections on consumer response to the concept and a product performance assumption, based on repeat profile for a new product in the category. This could be said as, “we are not assuming repeat rate, we are assuming product acceptance and use concept frequency as PC proxy.”
More: We also consider in-market data, previous testing experience, and marketing (especially timing) which can impact repeat.
Although FSIs can be cited as a source of awareness, they often tends to be low quality awareness,  decaying quickly.  FSIs can act as a consumer incentive generating trial, but we do not double count them as providing consumer incentive and driving consumer awareness.
Example: 1 FSI drop of 40MM circulation (Consider quick math)  
Redemptions roughly 1-2% in the US + A few additional % households see FSI and don’t redeem.  =  Estimate 5% catch attention: Roughly 2MM circulation of 116MM HH – less than 2 pts awareness.
BASES evaluates the volume impact of each shelf dynamic separately. These include store location confusion, package visibility, location on the shelf (top/middle/bottom), and product line space compared to competitors. BASES’ model also has flexibility to assess different shelf sets by outlet.
More: Simulated store tests combine these variables, which limits their ability to diagnose and suggest shelf solutions.  The product is always included in the competitive set, therefore it can’t evaluate store location confusion, which is important in new product launches.
Consulting Tips - - Alternate Nielsen Solutions: FindTime 
Applicability Does not apply to Pharma.  Does not apply to Pharma. Only for NABU and EMEA countries with coupons.  Does not apply to Pharma. 
  Question 31 Question 32 Question 33
  In your franchise growth model, when we say we will lose 3 SKUs to make room for the line extension, are you assuming that we are losing “average” SKUs? I don't believe the FGA results - these are inconsistent with how we view sourcing, and how we treat sourcing in our financial models. The BASES model is very US based. How do you handle the impact of the heavy promotion (twin packs) and the use of event promoters (push girls)?
Answer The FGA model is an empirical model. We do not assume you lose “average SKUs.” In-market history suggests that retailers remove worst performers to make room for new SKUs. We assume the SKUs removed are below-average performers. If you have specific SKUs you will remove to make room for the LX, we can incorporate those inputs into the FGA estimate. Consumer sourcing is included in our FGA model, however sourcing is only a partial view of true cannibalization effects, reflecting consumer responses alone. The franchise growth estimate also includes factors such as extent the line extension will borrow marketing support (e.g., spending, distribution) from the parent brand and differences in transaction size. 
If an approach other than FGA has been used: If historically a different approach (than FGA) was used to reflect cannibalization in your financial models, we can run calibrations using FGA vs. your method on historical cases.  I would not recommend switching outright.  The key to the treatment of incrementality in financial models used is applying CONSISTENCY across projects, and vs. historical cases.
The model includes country calibration to account for country specific promotions.
  • LatAm Answer: I would have agreed years ago, when we had less than 1,000 (Latam) concepts in the DB and less than 20 validations in the region. Today we have 6,000 concepts and almost 60 validations.  We have learned how to simulate non US-based promotions, like supermarket push girls, in-store product sampling, clowns handing out cereal samples to kids and good-looking girls sampling beer to guys, etc. 
  • Asia Answer: We think about all factors related to promotions and incorporate their impact. Some are model “enhancements” (like increasing DBA due to push-girls), while others are related to study design (adding promotional effect in concept board) for twin-packs. 
Consulting Tips Alternate Nielsen Solutions: Consider Nielsen optimization services, BASES DecisionPoint or Analytic Consulting Assortman.  - Consulting Tip: Ask the client questions to understand the promotional plan to model it. (e.g. what types of promotions, how many will you have, how many people will you reach, have you done a similar promotion before?, etc.) 
Applicability - - Does not apply to Pharma.
For LatAm and Asia-Pacific countries only. 

Forecasting Accuracy

Innovative Products are critical to client growth. BASES has a historical record of analysing all types of products including highly innovative products.

Contact BASES Product Support for the most recent category and country validation figures. 

  Question 34
  My product is totally innovative. Consumers won't give you accurate measures and therefore you will not be able to forecast it properly.
Answer Using the Success framework, truly innovative products would likely score outstanding on Distinct Proposition and possibly Attention Catching. Other factors indicate if consumers understand the product or identify a use for it. The specific factors help craft the message ensuring maximum potential for “new to world products.”
The BASES model was designed for “new to the world” products. Many competitors use a share-based model, which is problematic for initiatives that don’t fit into a category. BASES’ model uses absolute consumer response (adjusted for overstatement) to build the estimate rather than a “share of purchases” for the test product. The BASES model was designed for highly innovative products and our validation record is similar for these as our overall record.
Note: Caution: Be careful when talking about share based models because BASES has one. Gen-3 is not appropriate for "new to world" categories.
Consulting Tips Ask the client “What can I explain better so you are comfortable with the forecast?” This question could indicate a client who would discredit you perhaps not in the room but afterward. 
Applicability -

Marketing Support

BASES combines consumer response to new product concepts and products with the manufacturer’s planned marketing support to produce an estimate of future sales.

  Question 35 Question 36 Question 37
  How can I use the BASES methodology to determine the optimum marketing spend, or can I?
Ask: Ask client what they are trying to understand and provide learning slides or details based on volume drivers.
You've indicated that this subgroup is our prime prospect/high volume potential. If our marketing plan excels at targeting them, what is the volume potential? Are certain dayparts more effective than others in your model?
Answer The BASES model uses activity rather than spending. We feel confident in the model’s ability to assess marketing plans with different approaches and different support levels, but the BASES model was not designed to be a marketing mix tool.
If possible, offer guidance using Marketing Plan Analyzer.
This depends on success reaching the target group in the new plan versus the original plan. All plans have some level of targeting, therefore we expect general results to be comparable. Yes, different TV dayparts have different effectiveness weights in our model, evidenced by media cost across dayparts. 100 Prime GRPs cost more than 100 Early AM GRPs. Effectiveness weights are confidential, but BASES has worked with ad agencies finding BASES effectiveness weights are very similar theirs.
Consulting Tips Alternate Nielsen Solutions: AAC offers a Marketing Mix solution for established products.  Consulting Tip: Consider the size of the target group and their % of category purchasing.  If the ad agency is in the audience, remember they are media mix experts. Respect that expertise. 
Applicability Does not apply to Pharma.  - -
  Question 38 Question 39 Question 40
  Is there a difference in effectiveness for :15 ads vs. :30 ads? Do you have any evidence for modeling them differently? We will be spending $20MM on advertising the parent brand in the year of the Line Extension launch. How is that accounted for in the LX forecast? You're going to tell me to spend more money in Year 2. Is that always the best strategy?
Answer Generally, :15 ads are modeled less effective than :30 ads. Research on :15 ads shows they tend to drive less awareness and are often less effective at persuading consumers to buy. However, if you have an idea that is easy to communicate (like Coke Fridge Packs), a :15 ad can be a great tactic, and BASES model is equipped to handle those cases.

As a general rule, line extension advertising can have an impact on parent brand sales. However, parent brand ad spending tends to have a fairly insignificant impact on new line extension sales.
Ask: Ask if parent brand spending is consistent with spending in recent years.

  • If the level is similar, it is unlikely there will be much impact. Consumers are responding to the LX in the context of the level of parent brand spend
  • If it is a large increase, there could be an impact on LX sales; however, that might be overshadowed by the fact a competitor (the parent brand) is increasing support. To some degree, FGA addresses the “halo” to the LX due to contraction/expansion of the parent brand
     
To grow a brand in Year 2, you need to increase distribution or increase marketing support. For most initiatives, at least half the incremental growth in Y2 comes from trial.
More: BASES found products that grow in volume in Y2 spend as much in Y2 as in Y1, maintain or increase distribution, have strong acceptance, and tend to cast a more narrow net in Y1.  Because most new products do not get as much support in Y2, it is common for new products to experience sales decline in their second year.
That said, Spending heavily in the out-years can have negative profit ramifications.  For products with an intentionally short product life, consider a "big bang" strategy - go for maximum volume in Y1, and then milk it.  Conversely, products with longer projected lives.
Consulting Tips If the client has copy tested the initiative, ask them to share the full results.  - -
Applicability - - -

Marketing Execution

Clients often want to understand how the impact of first to launch vs. a pre-emptive launch by competitors, Marketing Execution questions address these concerns.

  Question 41 Question 42
  Give us perspective on the volumetric impact if we’re not first to market after all. What is the volumetric impact if a competitor increased spending in response to my launch?
Answer

Generally earlier launches into the category have higher unit and retail volume. First-in products also tend to experience higher and faster distribution, enjoy more trade promotion and are generally supported by heavier advertising spending.
If you are second entrant: However, you can overcome being second entrant. It is better to be the second entrant by launching with a great product (rather than beating competitors with a sub-par product.) If you are second to market, spend on advertising, have a stronger product and leverage parent brand equity with trade (fast distribution, in-store support).
 

Would you expect a higher response from a competitor than what’s typical? The BASES model is built from actual in-market launches. Most new products will be subject to response from competitors therefore, competitor response is naturally built into the model. The degree of impact is a function of amount of competitive response, category price sensitivity, uniqueness of the test products’ benefits, etc. If competitive response is particularly strong it may have a negative impact on test product volume.
Consulting Tips Alternate Nielsen Solutions: Recommend a future state or interaction design to test the impact of being second to market. Or, DecisionPoint depending on objectives.  Alternate Nielsen Solutions: Decision Point could be used to evaluate this situation. 

Holder and Refill Initiatives

Holder and Refill Initiatives are products which are dependent on each other for success. Understanding the relationship provides context of how consumers purchase these items. 

  Question 43
  If people are spending $10+ on a starter kit why isn’t repeat 100%? (reported repeat is in the high 30's)
Answer

While Holder/Refill initiatives have higher repeat rates than non-holder refill, expecting 100% is unrealistic. Why?

  • Consider purchase frequency. Some buyers haven’t had the opportunity to make a repeat purchase because their trial purchase was a near the end of Y1
  • Determine if the product performance meeting trier expectations
     
Consulting Tip -
Applicability Does not apply to Pharma. 

Analyst Challenge Summary

Analyst Challenge Summary

The Analyst Challenge training provides a wealth of details on challenging client questions and answers. Now you should be prepared for the most difficult situations with your clients.

This training resource is available for you to utilize as needed to review these questions and answers. Please discuss more in depth questions with your peers and management.