Faster, Leaner, More Predictive: A Modern Innovation Stage-Gate Process

Many large CPG firms are struggling to innovate fast enough, and this is leading some to experiment with different business models and different ways to get things through their large systems. The cry of ‘be Agile!’ can be heard in many corporate offices, and with this push to be more agile it could be easy to think that stage gate processes should be abolished altogether.   

However, before your team goes rushing to burn down the halls of the stage gate process, let’s take a step back and discuss why it was set up in the first place – and following that – how you can make it faster, leaner, and more predictive – without entirely abandoning the positive things a stage gate process provides.

Why have a stage gate process in the first place?

The goal of any stage gate process is to increase the ROI (Return on Investment) from the company’s spend on Innovation and New Product Development.

How do you achieve higher ROI? In my role at PepsiCo where I helped define and standardise the research steps within the stage gate process, there was a function called ‘Commercialisation’.  Led by Garry Clancy, one of the primary responsibilities of this function was to manage and run the stage gate process – of which research is just one input. Garry had a simple way to explain how we increase innovation ROI:

“Do the right projects, and do the projects right”

Garry Clancy, Senior Vice President, Global Innovation & Commercialisation, PepsiCo

So what does that practically mean, and why is it important?  If you are a small company and you only launch one or two innovations a year, then you don’t really need a stage gate process.  Setting up an agile team and having regular check-ins with stakeholders will enable you to have the right level of input to avoid risks and optimise your ideas. You can still use the research tools I will describe below to improve your ROI, but you won’t need a stage gate process with the same level of formality and oversight as a larger organisation.

However, if your company is launching multiple innovations each year – potentially across different divisions or regions, then it becomes more important to have a standardised system for evaluating innovation and comparing innovations against each other.  Decisions need to be made to prioritise resources and ensure that the right ideas are being worked on (those with the highest chance of a strong ROI), and that those ideas are as good as they can be (fully optimised).  And this is exactly what the stage gate process is designed to do.

Are all innovations equal?

No!  Not all innovations are equal in that they don’t all require the same amount of Investment to create and launch. Nor do they all carry the same potential in revenue.  So the potential ROI is very different from one innovation to the next. So too then the flow through the stage gate process must be different for different types of innovation. 

There are many different ways to classify innovation types, and each really depends on the context of your company. The basic axis to operate on is Risk/Reward. What is ‘high’ risk/reward innovation in one company might be considered ‘medium’ or ‘low’ in another.  But to help move the conversation forward, here is a simple segmentation of innovation types. Each of the following types of innovation would have different requirements as part of the Stage Gate Process.

Risk/Reward Levels

“High”

• Breakthrough innovation – requiring large CAPEX spend (i.e. a new product form), or entry into a new market, or a major communicated reformulation/restage of a large existing product

“Medium”

• New sub-brand, a medium spend in Promotion and/or CAPEX, or a restage of an existing medium sized product/brand

“Low”

• A flavour or variant line extension with minimal CAPEX or new ad spend

The Stages and Gates

For the purposes of the rest of this article, let’s assume that you are a large CPG firm that produces products that are sold on shelves in Supermarkets. The number of gates and the type of decisions made at each are the same regardless of whether your innovation is High, Medium, or Low.

What changes is the level of rigour that is needed to pass each of the gates, with higher risk/reward innovations needing more rigour & research, and innovations that are lower requiring less rigour/research.

Gate 0 – Strategy

Gate 0 isn’t really a gate, but it’s an important part of the process.  Gate 0 is all about Strategy, and the research tools that help you discover white spaces – find areas of potential competitive advantage, understand the needs and wants of your potential customers and ultimately help your company decide ‘Where to Play?’ to drive the most growth. 

Aligning the organisation to the output of the research and strategic thinking work done in Gate 0 will pay many dividends in the rest of the process – and is the foundation of the ‘Do the Right Projects’ part of the equation.  In fact – high failure rates in subsequent gates is often a symptom of not spending enough time/resources on getting Gate 0 right.

The goal then of Gate 0 is to understand the current and future category trends in as much detail as possible – and then to take that understanding and align it to your company’s capabilities and business strategy. The research should help you answer questions like: What product types are consumers gravitating towards? What claims/product attributes are resonating?  What new brands are gaining traction and why? What consumer behaviour is driving the change?  What are the important trends not just now, but in the future? What part of the market should I target? There are many different types of research methods, tools and companies that provide for these questions. Four common types:

• Segmentations
• Usage & Attitude Studies
• Category Trend Studies
• Ethnographies/Online Communities/Focus Groups

Each of these types of studies can be conducted in both traditional and non-traditional (next generation) ways and there are many companies that offer the above, from our own Black Swan Data to the big full-service providers like Kantar and Ipsos.

Things to watch out for:

• Does the research provide a longitudinal view, or is it just a snapshot in time?
• Does the research provide a prediction of the future?
• Does the research capture all of the trends in a category, and help you find those ‘unknown unknowns’?
• Does the research help you prioritise and make decisions, or does it just give you a list of things for further investigation?
• Does the research give you enough practical detail to action immediately, or are the findings too high level/generic?

As a final note on Gate 0 – although all of the gates require stakeholder alignment and buy-in, it is MOST important in Gate 0. If the organisation doesn’t buy into the strategy or the reasoning and research behind that strategy, getting ideas through the rest of the gates will be much MUCH harder. So be sure to spend the time here on internal buy in.

Think of it like a product launch to consumers – and treat it such inside your organisation. Market the hell out of it. Get T-shirts, mugs, posters, videos created, whatever it is that will help embed the ideas and get everyone talking in the same language about the segments in the market and which ones the company is deciding to go after.

Gate 1 – Idea

Stage 1 (which ends in Gate 1), is all about Ideation.  Let’s start off by defining what an idea is, and how it differs from a concept.  An idea should have just enough information in it to get the main consumer relevant points of the new product across – without tying yourself down to too many executional elements.  Concepts, which come later on in the process, build on these ideas and add in more executional detail like a Pack Image, variant line up, and even a price point/pack size/weight.  Different testing systems (more on this below) have different standards, but in general these are the elements of an idea vs a concept

While it is true that ideas can come from anywhere, and they should certainly be welcomed and captured, ideally, they would start from what you learned and aligned on from Gate 0 – i.e. ideas that fit the consumer needs identified in the market whitespace you chose from Gate 0.

In the Idea phase your company should be generating 10’s to 100’s of ideas – and then finding ways to screen through them as quickly and objectively as you can – with the goal of using consumer feedback to help you provide an early read on in-market viability. Try not to screen out too many ideas at this stage that you don’t currently have an R&D capability for – if it fits the needs identified in Gate 0, there are many ways to acquire the capability needed to produce the product.

As you are generating these ideas and preparing them to get consumer feedback, it’s often unavoidable to have extra ideas thrown in that are not on strategy – common sources are executive ideas (HiPPO – Highest Paid Person’s Opinion), R&D capabilities searching for a home and a use case, or things that competitors are doing that aren’t in the market whitespace you have already identified. 

Do your best not to include these – as even if they come up with strong results, they could lead you off strategy which has other implications. If you can’t avoid it – try to remind the stakeholder of the agreements you made in Gate 0 and get them to self-identify that their idea might not be on strategy.

When it comes to getting consumer feedback to help you objectively screen and prioritise ideas, there are many ways to do it including:

1) use a survey to get quantitative analysis

2) run a focus group and assess them qualitatively

3) use a big data tool that uses secondary data like we do at Black Swan Data

Options 1 and 2 require showing a consumer a stimulus, and prompting for feedback to that stimulus – in option 1 in the form of a survey, in option 2 in the form of a live conversation.  For a more in-depth review of some of the challenges of these primary research based method, see my previous article here: Is it the beginning of the end of Primary Research?

If you move forward with either option 1 or 2, a few things to ensure you keep in mind. Primarily, that the key concept behind using surveys or qual to screen ideas is benchmarking.  No matter what the response is – positive or negative, the only way to know if that response is important or significant is to do three things:

Compare the results you have for that idea vs the results you have from other ideas

Ensure all of the variables EXCEPT the idea stimulus have remained the same across studies – I.e. the sample composition, the questions asked, the question order, the collection methodology (online, offline, mobile etc), the order the ideas are shown in, even the format of the stimulus (images included or not, where text is placed, etc)

Ensure you have enough ideas to compare against (in a test vs a database) and that those ideas represent a spread of good, bad, and average ideas that are relevant for the category. Usually you are looking for 30+ ideas to bench against – the more the better.

If you go with a more traditional technique like option 1 or 2, here are some companies to look at. They generally vary along 3 axis:

• The size and comprehensiveness of their benchmark database
• The speed / cost of the study
• Whether you set up the study yourself and analyse it (DIY), or have someone else do it for you.

rd option involves using data that already exists to validate that your idea has consumer potential – enabling you to replace what traditionally could only have been done by getting new primary research data via a survey or focus group.  This is a new and exciting area of innovation in the market research space, and one that we at Black Swan Data have been leading the charge in for the past few years.

We track online conversation about hundreds of thousands of trends, organised into category datasets.  Our tool, TRENDSCOPE, allows clients to build ideas based on what consumers are naturally talking about – both about a specific product like Matcha Tea or Alcoholic Seltzers, but importantly also about other connected trends, like which benefits or ingredients consumers are talking about together with a specific product.  This allows our clients to build new ideas quickly and confidently without having to run an expensive or time-consuming survey – using the volume and predicted volume of consumer conversation to prioritise the ideas with the best chance of future success. 

Whichever option you choose, at the end of Gate 1 – the goal is to have found ideas that meet a minimum threshold of consumer appeal, and ideally also a decent amount of uniqueness (although this may be less important for some innovations). 

Alongside the validation from consumers, internally you are likely to have estimates about the amount of investment/CAPEX it would take to actually produce the new product described in the idea and the combination of those two things – essentially value vs. effort.  The ideas that pass this gate form your ‘hopper’ of ideas that are ready to progress to the next stage should be in a priority order based on the value vs. effort, and also a strategic assessment of fit to strategy.

Gate 2 – Concept

Work done in this stage is all about refining and building out the full concept. In this phase goals – and how much money you spend on researching these goals vs making a judgement call  – should depend on the level of Risk/Reward of the Innovation (High/Med/Low as described above). 

Here are some typical goals:

⁃ Choose and optimise the ideal pack design

⁃ Choose an ideal new product name

⁃ Choose the most relevant expression of the key benefit

⁃ Choose the ideal product variant/flavour line up

⁃ Choose and optimise the price point

⁃ Validate the overall concept

⁃ Forecast the potential revenue from the overall concept

All of the points here are conceptual – no actual product development has begun yet, and there are qualitative and quantitative ways to achieve most of the goals above.  Many companies use a combination of quant/qual methods -starting with qual to get some initial feedback on things like the benefit expression or the pack design, and then moving to quant to choose and validate the decision. 

For the quant part, nearly all of your options will be survey based. Traditional methodologies include: monadic concept testing, conjoint, TURF/SURF analysis. Newer methods include: Competitive Markets (similar to chip game studies), Eye-tracking for Pack Design tests, or the genetic algorithm based method of Affinova (now part of Nielsen/BASES).

Nearly all full service market research firms will offer all or most of the services and methodologies above, but here it starts to pay to work with more specialist vendors. That is because as we said in Gate 1 – if you are using a survey based method to validate ideas – what really matters is WHAT you are validating them against, I.e. what are the benchmarks the company has to compare your ideas or new concepts against. Some of the newer methods like competitive markets claim that they don’t need databases of previous studies and can rely on the comparison in a single study – but even those require that the concepts you put into that single study are all relevant for the category and represent a good spread of ideas.

In addition, many of the solutions above are available in a fully customised form from the vendor, but also now many are offering standardised versions on DIY platforms like Zappi or on their own similar platform.  Some vendors to check out for this phase:

• The traditional heavyweights – BASES (Nielsen IQ), Ipsos
• Competitive Markets – System 1, Veylinx, Hotspex
• Pack Testing – PRS, EyeSee, Dragonfly.ai
• DIY Platform – Zappi

Black Swan’s TRENDSCOPE can also help pair with many of these services by helping you screen down options like ideal flavours, or benefits before going into costly validation tests with the above vendors.  In tests we’ve done with some of our clients we’ve seen significant uplifts in concept scores by optimising elements of their concept using our TPV (Trend Prediction Value) and Associations metric to find emerging elements with higher relevance and uniqueness. So while the final validation is still done by one of the vendors listed above, you can both get better scores and reduce the number of concepts you have to test if you use a refinement/optimisation tool like Black Swan to help you better prepare. See case study here for more: How Trendscope generated $64m ROI through Concept Optimisation.

At the end of Gate 2, the goal is to have consumer validation of your full concept proposition – i.e. a specific product, under a specific brand, with a specific benefit and reason to believe, with a chosen price point, flavour line up, and pack design.  Essentially, everything you would see if you were to walk up to that new product on a shelf in a retail store.  Passing the gate means you have achieved a sufficient score against the benchmarks from the concept test on things like Purchase Intent, Value, and Uniqueness and also your internal finance and R&D teams will have started doing some modelling on CAPEX investment needed to create and launch the new product. The combination of those two things: sufficient consumer validation, and sufficient internal financial modelling/payback is what you need to pass Gate 2.

One last note on Gate 2, it can be tempting for marketers to ‘stack the deck’ in a concept in order to pass this stage – making claims in the concept that will be hard for the product to live up to in reality.  Doing so will certainly improve your concept score – but remember the stage gate process isn’t always linear. If in the next phase the R&D team isn’t able to create a product that lives up to the overly inflated expectations set by the concept, then you’ll need to revise the concept, re-test and come back to pass Gate 2 again, losing both time and money as you rework the concept. Not to mention pissing off your colleagues in R&D…

Gate 3 – Concept/Product Fit

The whole goal of this phase is to achieve concept/product fit and will be heavily led by the R&D team.  No product is objectively good or bad independent of an expectation.  Our human brains place an expectation on everything we see before we experience it – and how good that experience is will depend on whether it met, exceeded, or failed our expectations. 

Even in a blind product test there are expectations consumers will have before they trial the product – we just don’t know what they are. So the goal here is to prime consumers with the appropriate context as they would get when they bought the product in-store (i.e. the concept you tested in Gate 2), and then give them the product to experience and see if it meets their expectations. 

Along the way the R&D, Insights, and Marketing teams may conduct any number of studies – both qualitative and quantitative to help them optimise aspects of the product. If it’s a food product there may be many ways to achieve the desired texture, or thickness, or flavour profile/intensity and many of the same techniques and vendors described in Gate 2 also can be used here – with the addition of having consumers actually try the product.

At the end of Gate 3 is the most important decision of the whole process – the ‘Go/No Go’ decision that is made before any significant CAPEX is spent. Here again we look at the Risk/Reward profile of the innovation, and the hurdles each has to go through at Gate 3 will be different. 

If it is low CAPEX – for example a limited edition flavour variant – very little will be needed in terms of validation.  However, if it will require $50-$100M in CAPEX to build a new factory and manufacturing process to create the new product, then you will likely need not only a full concept/product test that passes all the benchmarks with flying colours – but also a forecast of Yr 1 and Yr 2 revenue.

The forecast can be received from an STM (Simulated Test Market) from companies like BASES/Nielsen IQ where they combine the consumer input from a concept/product test with the marketing/launch plan you give them (i.e. how much distribution will be achieved, how much promotion and advertising spend there will be, etc) to predict Y1 and Y2 sales for the new product launch.  These tests are expensive so only worth doing on your highest risk/reward innovations – but the test is far less expensive than wasting $100M on a new factory that goes under-utilised.

My personal point of view on STM’s is they are useful for high risk/high CAPEX decisions, but they are generally too expensive and slow – and most are optimised for a predominately offline/bricks & mortar world.  As more and more shopping goes online, these methodologies will struggle to retain their accuracy – and new methodologies will arise that are more predictive, cheaper, and faster. I talk about one potential method here: Is it the beginning of the end of Primary Research?

Gate 4 – Launch

At this point you have the green light to go ahead, and the Supply Chain function are busy building the factory and equipment needed to produce the new product. The focus here now shifts to launch preparation with the Marketing and Insights team focusing on distribution, promotion and communication plans. 

Comm’s is a critical component of every new launch – and will play a huge role in setting consumer expectations for your new product.  Working in partnership with your advertising agency, you’ll likely go through a process of qualitative research -> quantitative research that in many ways is similar to the process in Gates 1 and 2 – with the difference being that here you are testing an advertising  idea vs the overall product concept.

Given the similarities with Gate 1 and 2, the methodologies you can use are also similar.  One key consideration here is to incorporate a better understanding of cultural trends beyond just your category. To help you ideate and screen ad ideas you can use companies like Sparks & Honey, Discover.ai, or Black Swan Data!  When you get to creating a script or animatic of your ad idea, similar challenges exist as in Gate 2 – i.e. if you want to validate that your advertising concept is good, you need to have a way of benchmarking that idea vs other ideas using the same method.  Again, here you can choose between fully custom tests from most full service research providers or you can employ standardised solutions from companies like Zappi.

Either way, if you are setting up your research standards for a stage gate process for the first time – try and settle on one methodology. It will make the comparison across innovations much easier, and you’ll only have to educate your organisation once on why they should believe the results.

At the end of Gate 4 is the final launch decision. By this point you should have a killer advertising campaign, strong retailer rollout plans, and R&D should have tested the final product off the full assembly line to ensure it tastes as good as the testing samples that were used in the Go/No Go decision at Gate 3.

Gate 5??? 

Hooray!  The innovation is out in the world and flying off the shelves. Time to move onto the next project, right?  Not exactly.  While the formal stage gate process is over at this point, I’d encourage companies to think about adding a ‘Gate 5’.  That Gate isn’t a gate in the traditional sense. It’s not a go/no go decision, but rather a formalised point of reflection conducted 12 months after launch to review performance.  Did you achieve the distribution you expected?  Did the marketing perform as well as projected?  Have sales hit forecast? The goal here is to understand what has happened, and use these learnings to improve future innovations (as well as input into the Year 2+ strategy of the innovation that was launched).

All of the Gates I’ve described above may seem like a lot – but remember how much rigour you apply at each stage should vary by Risk/Reward. For the highest Risk Reward Innovations you may do several studies at all of the gates.  For the lowest – you might only do two studies for the entire process – one to check the concept, and one to check concept/product fit. The process can be as fast or slow as needed – but the important thing is it gives you a common framework in your organisation to make objective data-backed decisions. 

Dave Soderberg is Black Swan Data’s Chief Data Officer.  Prior to working at Black Swan, Dave spent over a dozen years working in agency and client-side market research roles for PepsiCo, Spectrum Health and BASES. To talk to Dave or a member of the team please contact us at [email protected].

The post appeared first on Black Swan Data.