Tuesday, April 10, 2018

Making Ads Part 3 - AD (After development)

I have been meaning to write this for a while now but then life took over & then this took a backseat. I needed a break from my craft as I have fondly started calling marketing. Well, this is the closest I will be able to call myself creative. So, I decided to take refuge in my second favourite passion - travel!!! This post comes to you from Japan. I am currently at this moment on my way from Osaka to Hiroshima - from modernity to history is what I am choosing to call this journey. It is also the sort of contrast I enjoy in storytelling & advertising. 

Till now on our journey of making ads we have covered the science behind making ads & how to nail a brief. From the release of the brief till you actually decide on the script & then shoot it there is a long period. I call it the dramatic pause. How all performing arts have this moment when you wait with your breath held for something big & dramatic to happen. This is the time the creative team is taking to come up with the brilliant script that will overwhelm you. 

This post is about how to marry whatever we have done in part 1 & part 2 with the actual creative work for bringing out at ad/campaign that works. Now first thing to be convinced about individually and sadly no one will be able to change your mind if you believe otherwise is that advertising’s sole purpose is to sell. It will sell better if its memorable & can break clutter but its purpose is still to sell. Its purpose is not to entertain or win awards. Well if awards is the purpose then the brief should be different and yes maybe once in a lifetime you will hit the jackpot and make advertising that will do both. I have not seen many & hence I will focus on what is the usual order - ads need to increase sales. Of course the route to more sales is better brand recall, better brand association, better brand love & eventually building propensity for trail. 

So where do we start with this part of ad making - After Development (AD). Ensure that the script/campaign is presented by the creative themselves in a meeting. Please do not get scripts or creatives emailed to you. So much is lost in translation when you read something on your own. The tone & the visualisatio  is completely lost when seen on a computer screen. Its like reading a book & making your own assumptions about what is the character like till someone else actually tells them your POV. This part of the process is also where your right brain needs to take over your left brain. You need to become a consumer. As Dumbledore tells Harry in the 6th book - from now on we will leave the known & venture into guesswork. Appreciating a script is pretty much like that. For every brief written millions of scripts can be written and you will never be able to tell which is better even on hindsight because you will choose only one and move forward. 

I will however try & outline some of the things that can help you make a better choice - its not fool proof but then most things in marketing are not. Unlike potion making that is exact; marketing is a field of magic that needs individuality & judgement (apologies to the non Harry Potter fans - I can’t help but draw analogies from there - I promise the post still makes complete sense. It may lack the fun for you :p)

1. Put yourself in the TGs shoes: evaluate the script/campaign as your TG will. You should know your TG inside out and when you hear he script/campaign think about what she/he will think about it. This is often easier said than done. This is where pop culture & mass entertainment comes to our rescue. Look around yourself for trends that the consumer is responding positively to & then see if your script/campaign comfirms to it. Here, be a little cautious. While it needs to confirm to the mass trends - it cannot sit in the center of it. It needs to sit just a tiny bit outside of it to be perceived as interesting & new. Eventually leading to breaking clutter. Example if you are showing a conventional husband wife relationship; it cannot depict a submissive - you are my lord wife. There needs to be a new age bend to it. 

2. Remember the brief: even if the script blows your mind; if it does not deliver to the brief - do not proceed. Ask the creative team the question - how does it deliver to the brief? We are all guilty of choosing the story over the logic & in a few & rare cases the punt also pays off. But more often than not, one winds up with a piece of creative that is great; overwhelmingly good but does little to do what it was supposed to - increase sales. 

3. Confirmity to brand codes: the script needs to sit within the brand world you have curated & carefully nurtured over years of existence. For a new brand this would be the world you need to weave & be clear about before you look at the creative. The brand world determines the creative language & not vice versa. Example if you are a brand based on the vibrancy of youth & everything it brings with itself - college life, need for new experiences, need for edginess all leading to a bit of flippancy or volatility 
(E.g. Fast Track); talking about changing the world & youth is the power of tomorrow (E.g. Tata Tea) is not being true to your brand world and even if its the trend & is being adopted across categories & pop culture - do not accept it. 

4. Brand is a part of the story: This is where you & your creative team will differ the most. But this is where you come in as the custodian of the brand. Your brand needs to be a part of the story. The ad should not be complete without the role your brand has to play in it. For example in the Asian Paints (har ghar kuch kehta hai) ad; the story is not complete without Asian Paints at all. Your brand needs to be the phoenix feather wand in Harry Potter - the story is not complete without it. A large part of this is aided by your brief & the brand world you create but the rest is the job of the creative. And yes sometimes your brand uniquely may not fit into the story but at least the category needs to. In large, highly penetrated categories with multiple brands straddling the same benefit segment to have the brand uniquely play a role in the story is unrealistic. But yes the category & most importantly the benefit segment definitely needs to. For example fresh breath needs to be integral to a toothpaste ad even if its not Close up. 

5. Let your heart respond: once you have done the first four its time to let the right brain take over. If the script/campaign does not speak to your heart then its not the right script. It needs to appeal to you as a creative piece of work much like a book or a movie or a song or a video or a TV show. And it doesn’t need to be just emotional - it needs to just be something that instinctively appeals to your right brain - could be emotional, funny, nostalgic, dramatic, shocking et al

With this I finish my discourse on ad making. There is a whole lot to be done as you move from here to actual production & releasing it. That part mostly is the operational bit of ad making. I personally love it - Pre-bids, treatment notes, PPMs, offlines et al. But I am not sure if anyone is interested enough to want to go through 
that process in detail. If there are people who want to know, leave a comment and if there is a quorum; we shall take you through that process as well. 

P.s. Apologies for typos if any - its been written on the phone in small print on a train. So please excuse. 

Sunday, April 1, 2018

Designing Pack Price Architecture (PPA) through Conjoint

Part 3/3 from Jayant - insightful piece on Price Pack Architecture to optimise portfolios. From whatever I've seen, this is one of the most effective parts of the Revenue Growth toolkit.

Designing Pack Price Architecture (PPA) through Conjoint

Most consumer decisions are made within the influence of several factors – price, competitive marketing efforts (media and trade promotions), brand significance, share of mind/shelf in the shop/retail, new innovative launches etc. So how can a researcher decide on a methodology that allows us to study the impact of these variables when optimizing a brand portfolio.
Price Pack Architecture is important to ensure that each SKU in a brand portfolio is aiding in maximizing volume and value for the parent brand (with least cannibalization). Since each SKU would differ slightly in its offering (size, price, flavor etc.), it needs to be understood what impact does it have on the business by changing or introducing new SKUs or brands in the portfolio.
Conjoint is one of the most commonly used research tools to study the impact of different PPA scenarios, e.g.
·       what would happen if we change the size (grams) of one of our SKUs?
·       Does a 10-cookie pack cannibalize my own 5 cookie pack or take from competition?
·       Does reducing price by $1 hep increase volumes enough to bring more revenue?

How it works?
Collecting data – The exercise is simple – each respondent is shown a series of scenarios with different brands and SKUs at different price points. These SKUs and price points vary for each scenario. We may show some brands and SKUs in some screen and not show in another. The same SKU can have a variation of price points (generally defined as +10% or +5% of base price).
Then for each scenario the respondent is asked to choose an option they would most likely buy (assuming this were a real purchase). To make the scenario true to reality we sometimes also offer an option of “none/buy something else” to allow the respondent an option to reject all options in a scenario. Below is a screen to illustrate a survey screen. Generally, a respondent is exposed to 10-15 such screens depending on the number of brands/SKUs we are testing.

Interpreting the output –

        Price sensitivity analysis is done to determine Optimal price points: Below is an example of a price sensitivity analysis – it shows that if everything in the market were to remain constant, how would our brands value and volume sales change if we alter the price. The slope of this curve is called price elasticity (i.e. the change in value/volume sales if we change the price by 1 unit). Brand managers desire to have price in-elastic products (i.e. very small price elasticity ~0). A low-price elasticity would indicate that the brand doesn’t lose sales as we increase price.

The price range to play in would be the region between two break points i.e. the points where there is a biggest drop in sales when price is changed

        Forecasting volumes in different scenarios – we can create scenarios and predict the volume/value sales changes. E.g. look at the example below –
        The portfolio mix with an addition of small pack (4 cookie) in Choco Crème leads to an overall 6% and 4% growth in volume and revenue respectively of the Oreo portfolio
        However, this portfolio mix leads to 7% decrease in both volume and revenue respectively for the Vanilla variants

Running different permutation and combination of such scenarios will then help estimate gain/loss to Oreo portfolio, thus deciding an optimal portfolio. We can also create scenarios where the competition prices and launches can be presumed to predict different “what if” scenarios.

KPIs for Brand Health Tracker: Reading BHT Metrics

Here's part 2/3 from Jayant - some basics in tracking your brands health

KPIs for Brand Health Tracker: Reading BHT Metrics

As a marketer tracking sales helps understand your market standing. But it doesn’t answer ‘why” your sales are the way they are. You may enjoy a large share but an undesirable brand strength or vice versa. Hence it is important to measure the health and equity of a brand in the market.

Brand health trackers are the mostly used “research tool” to track (hence the name tracker) and measure the standing of brand and its competitors in the market. It is a standard report that is generated on a consistent frequency. This frequency may vary from industry to industry (e.g. it’s important for CPG brands to measure key brand health metrics on monthly or even weekly basis).
Each such iteration is generally called a wave, so you read movements of key brand health metrics wave on wave. E.g. If we do brand health tracking twice a year, then you have two waves a year. 
Generally, in practice you wish to see movements of data for last 2 to 4 waves depending how far back in history to you wish to go.
In this short write up we will go through some of these key brand health metrics and analysis, which you should know as a brand manager –

Brand Funnels: As shown below, this funnels helps us understand how good the brand is doing at all aspects of customer decision making process. E.g. below graph shows us that our brand is performing poorly on translating consideration to trial and making sure we are the Most Often Used Brand (MOUB). Other brands are performing better at these aspects.
·       Awareness – The union of unaided and aided recall gives us ‘awareness’. Being low on this indicated that the brand needs to increase share of voice in the media/communications – more campaigns and communications to help improve the awareness of the brand
o   We ask “which brands are you aware of”, followed by an open-end text box. %age of respondents mentioning the brand gives us “unaided recall” of the brand
o   After which we can show a list of brands and ask “which ones of these brands are you aware of”, %age of respondents mentioning the brand gives us “aided recall” of the brand
·       Consideration – Within the brands that the respondent is aware of, it is the percentage of respondents who say they have considered purchasing/using the brand in the past Being low on this indicates our positioning or offering isn’t right i.e. consumers do not think of us as a valid choice to fulfill their needs
·       Trial – Within the brands that the respondent has considered, it is the percentage of respondents who say they have purchased/used the brand in the past Low on this generally indicates that our availability or price is an issue. Because what is
·       Usage – Within the brands that the respondent has tried, it is the percentage of respondents who say they are currently using the brand (in case of categories like confectionary, services like restaurant, hotels etc. we ask if they have used it in recently e.g. last 1 week/month etc.)
·       Most often used brand (MOUB or BUMO, Brand used most often) – Within the brands that the respondent is currently using, it is the percentage of respondents who say that the brand is their most often used brand.
Very important: For consideration, trial, usage and MOUB we generally ask respondents to not think beyond 3 months for CPG brands and up to a year for high involvement categories like cars, appliances, hotels, vacations etc.

Share of Heart/ Share of Mind: The Share of our brand versus other brands in the unaided awareness is sometimes termed as “share of mind”. Similarly, the share of our brand versus other brands for MOUB is often referred to as “share of heart”.
Likelihood to used again – Sometime, post the funnel questions, we ask respondent how likely are they to continue using their MOUB. This metric is asked on a 5/7 or 10-point scale. The percentage of respondent answering the top score or top 2 scores are reported as the loyalty metric.  
NPS – In the service sector (e.g. hospitality, telecom, software etc.) there is standard metric used called Net Promoter Score (NPS). Respondents are asked how likely are they to recommend a brand. This is asked for our and competitor brand on a 11-point scale (0 to 10). Percentage of respondents giving top 3 ratings are called promoters and bottom and bottom 6 ratings (0 to 5) are called detractors. Promoters % minus detractor % gives us the NPS score. There are other ways to calculate this and it is a topic that warrants a detailed discussion. So, till then we will pause here.
Brand Imagery: This is the second most important part of a brand health tracker. For each brand that the respondent is aware of OR have used in the past, we ask to rate the brands on different attributes.

Attribute List: Below are some examples of attributes that a brand can track –
This brand's products taste great
This brand Helps me be independent
This brand has packaging I like
This brand helps me to keep doing the things I love
This brand comes in Flavors I Like
This is a brand I trust
This brand is widely available
This brand’s products offer good value for money
This brand is Easy to find in the store
This brand’s products are of high quality
I enjoy watching the ads from this brand
This brand brings new and innovative products
This is a brand I love
Is a premium brand
This Is a brand I want to be seen using
Is an inspiring brand
This Is a popular brand
The brand is a leader in the market
This brand's products have a texture I like
This brand's products are different from other brands
Is on sale often
This brand is recommended by my friends and family

Some themes of attributes are –

Functional Benefits (price, value, ease of use etc,)
Brand positioning (emotional in nature e.g. trust, brand for me, scientific, cool, modern etc.)
Category Leadership
Recommendations/ Endorsements (reco by friend, reco by expert, reco by salesperson)
Retail Availability

Owned versus Opportunities: When comparing the brands across different attributes, you will notice attributes which our brand is the strongest versus competitive brands. These are ‘owned’ attributes, some attributes may be owned by competition as well. There is also a low hanging fruit – attributes which are not owned by any brand – i.e. all brands have a low rating on these attributes. These low hanging fruits and competition owned attributes are the opportunities for our brand to act on.
Ad/Message Recall: Sometimes brands wish to measure Ad recall along with brand health tracker. The most common ad recall metric is –  Rate of Recall: Percentage of respondents who mentioned they recall seeing an ad by the brand. This again can be asked as aided and unaided, just like the awareness metric. We also follow up this question by asking where do they recall seeing the ad – TV, print, POS, radio etc.
Purchase/Usage behavior: Brand health trackers also allow opportunity to ask behavioral and usage data. This is done especially for categories with growing distribution in the markets. Asking respondents where (point of purchase), when (how frequently) and how much (quantity) of product do they purchase.
Some Guidance tips:
·       Target population: This is important, because it will decide ‘who” will answer your brand health tracker questionnaire. Going beyond your target population will give us bad or useless data. e.g. the target for Horlicks would be Moms of kids aged 2 to 10 year old kids, from socio economic class A. B and C. these moms should be using or willing to use nutritional supplements for their kids.
·       Rolling data Versus Stand in Time: We measure brand health metrics either per wave, or we combine data for past waves to reduce any sample bias i.e. sudden jumps and movements (this is called rolling data e.g. past 3 waves rolled as one reading)
·       Brand Power: many research providers have their own trademark methodologies to give one single score brand equity score/measure.
·       Social Media: This is one aspect that warrants a special discussion. Generally, the social media indicators are ahead of its time than overall market, as early adopters are the first ones to talk about the brand or product online versus others. Also, social media analysis complements a tracker, it is still far away from replacing the traditional brand health tracker.

Being smart about research

A key area to nail to succeed as a marketer is research - it forms the basis of the decisions that you take. There are several areas to be explored in MR, so as a starter I requested Jayant to help clarify some  basic MR concepts, and introduce a few key frameworks (in the following posts).

I know Jayant from MICA, and he's spent the better part of this decade working in research, helping consumer clients form better decisions from their data at AbsolutData in Chicago. You can find him here.

Here's the first of three parts: Being smart about research:

Most marketers would have support from their Insights and Analytics team. These teams could work parallel feeding strategic insights at regular intervals or at times respond to Adhoc requests from marketing teams. In majority of the organizations insights and analytics team is an integral part of the marketing team itself, working closely on a day-to-day basis. A successful marketer would always be a smart researcher as well. This does not imply to have deep knowledge of techniques and methodologies but a quick understanding of three aspects –

1.       Explaining a Brief: How do I frame my question to get the best answer?

For illustration purposes, let’s us say the question is “Locate the positioning of a new brand line extension in the market for brand X”. This statement can be further dissected into three questions:
·       Current Brand Saliency: What does brand X stand for? My extension can’t be too far away from it?
·       White Spaces: What are some white spaces around brand X that this new line extension will target?
·       Immediate Line Extension Positioning Opportunities: How can I prioritize these different white spaces for my new line extension?

Now see below how these questions are translated into simple requests for the research brief
Jargon – What you speak?
Plain Spoken – what you actually need?
What you do research wise?
Current Brand Saliency
What is the perception of brand X in the market?
See on which brand attributes does brand X lead against competition
White Spaces
Which attributes are currently not owned by any brand?
 Identify attributes not owned by any brand (i.e. all brands have 50% or lower association)
Immediate Line Extension Positioning Opportunities
What is the perception of brand X on the “white spaces”
Locate which attributes the line extension can position itself on

2.       Suggesting methodologies: What scale/scope of research do I need?

We will cover these topics in details one by one in later posts. Here is quick cheat sheet of different methodologies for the key business questions you may be thinking about

Existing SKUs/ Brands
Nielsen data: Look at the Average Selling price versus competition
Market mix Modelling: See the impact of price changes: Price sensitivity graphs
Price Pack Architecture: Price sensitivity graphs
A/B Testing (this is mostly in Ecommerce domain)
New Product
Price Sensitivity Monitor
Gabor Grangers method
Conjoint (Optimization)
Profit Pool Analysis
Logo/ Design
Qualitative: Concept/Copy Testing
Quantitative: Concept/Copy Testing
Features/ Attributes
Usage and Attitude Study (also called A&U, U&A etc.)
Habits and Practices Study (e.g. diary based, sometime through panels etc.)
Trade Off Exercise: Conjoint (Optimization)
Promotions/ Campaigns
New campaign
Qualitative: Concept/Copy Testing
A/B Testing (this is mostly in Ecommerce domain)
Conjoint (Optimization)
Existing campaign
Market Mix Modelling
Ad tracking
LINK test (there are several such trademark studies available through traditional research providers)
Lift Calculation (Mostly in the tech/digital and E-Commerce domain)
Medium/ Platform (TV, Digital, OOH etc.)
Market Mix Modelling
Attribution Modelling (Mostly in the tech/digital and E-Commerce domain)
Place/ Medium
Traditional Channels
Channel Optimization
Market Mix Modelling
Nielsen Dashboards: Nielsen data can provide time series growth and distribution metrics
Attribution Modelling (Mostly in the tech/digital and E-Commerce domain)

You will notice that segmentation isn’t mentioned here – because it’s an expensive and elaborate exercise to do. You don’t sound smart by suggesting segmentation as a first solution to a business problem. Any brand manager would have access to the profiles of key segments in the market. A good segmentation is valid for 2-4 years depending how mature the market is. The more mature/stable the market, the longer segmentation is valid (e.g. US/Europe are mature markets versus Vietnam/India which are emerging markets).

3.       Reading data: Know how to read data?

Reading data across different types of consumers is called ‘Cuts’ or ‘Banners’. Some key cuts that you should always keep in mind are as follows –
Key Consumer Segments
Brand Users
All Respondents
Seg A
Seg B
Seg C
Own Brand
Comp 1
Age, gender, ethnicity, region, etc.
New Users
Loyal Users

A.      Total number is generally misleading, e.g.
·       There are consumers aged 18-20 yr and then there are some older ones at 50-55 yr
·       The average age would be 30 yr
·       In reality – there isn’t any consumer aged 30, so reading averages is dangerous
B.       Be careful in reading demographic data. E.g. Reading across ethnicity is pretty common in US, while it is unethical in EU
C.       Reading across consumer segments is important, as brands take decisions by keeping in mind only the “Key Strategic Segment”, remember – you can’t target all