Посты с тэгом: attribute dependency

The Three Faces of Attribute Dependency

Published date: October 26, 2015 в 3:00 am

Written by:

Category: Uncategorized

Tags: ,,,,

When using the Attribute Dependency technique, you’ll reach a point in the function follows form process where it’s time to make adaptations to your concept. That’s where you try to improve the concept and put more definition around it.
One way to make adaptations with Attribute Dependency is to change the type of dependency. There are three ways to do it: passive, active and automatic. Think of these as what has to happen within the product or service for the dependency to take place. Let’s look at each type.
Passive dependencies, just as the name implies, are passive. Nothing has to happen for the dependency to take place. There doesn’t need to be an intervening element to cause the dependency.
Look around and you will see that many products and services are examples of passive dependency. Here is a simple example of mixing bowls that come in different sizes.
Now you may ask, “Is this really an example of the attribute dependency pattern?” It certainly is. As one thing changes another thing changes. In this case, as the needs of the user change, the size of the bowl changes. It’s a passive dependency, though, because the bowls simply exist in various sizes and shapes. In fact, any product that comes in different sizes such as clothing, hardware items, even homes are examples of passive attribute dependency.
But some dependencies require an active, intervening element to cause them to occur. A very simple example is Happy Hour, when the price of drinks in a bar is reduced. But for this to happen, somebody has to do something. That active element, of course, is the bartender. At the appointed happy hour, let’s say 5 o’clock, the bartender simply lowers the price of the drinks presumably for an hour. Then again at 6 o’clock, the bartender raises those prices back to their normal level. Because of the active intervention, we call this an active dependency.
TransitionAnd finally, we have automatic dependencies. These are unique because they happen, as the name implies, automatically. The product or service is designed so that as one thing changes, the product automatically changes by itself without some intervening third-party element to make that change.
Transition sunglasses are one of the best examples of an automatic dependency. As the brightness of the light changes, the lens automatically darkens in response to that change.
Products that have this type of dependency seem almost smart. They know when it’s appropriate to change in response to some other variable, either an internal or external. The consumer doesn’t have to do anything because the product does it all by itself.
How do you know which type of dependency to use? It depends on a lot of factors such as how much convenience you want to deliver to the customer. Is it technically feasible to create a particular dependency? For example, your engineers might be able to make a mixing bowl that automatically expands as you put more things in it. But that also adds a lot of cost and complexity. It’s probably a lot easier for the customer just to grab the right size bowl to make a cake.
It also depends on how much control you may need in a situation. Do you want the customer or another person making the change? Look back at the happy hour example. You could create a cash register that automatically adjusts the price of drinks based on the time of day. The bartender wouldn’t have to think about. You would have complete control over the prices throughout the day.
Passive, active, and automatic. That’s three ways to give your customers very cool products with the Attribute Dependency technique.

Innovation Sighting: Adjustable Airline Seats

Published date: August 17, 2015 в 3:00 am

Written by:

Category: Uncategorized

Tags: ,,,,

Here’s a nice example of the Attribute Dependency Technique, one of five in the innovation method called Systematic Inventive Thinking (SIT). It’s a great tool to make products and services that are “smart.” They adjust and learn, then adapt their performance to suit the needs of the user. Attribute Dependency accounts for the majority of innovative products and services, according to research conducted by my co-author, Dr. Jacob Goldenberg.
From Fox News:

The airline legroom wars may finally be coming to an end.
Engineering firm B/E Aerospace has filed a patent for a “legroom adjustable” seat design that allows flight attendants to move a seat forward or back depending on the size of a passenger, reports the Telegraph.
The seats, which all have moveable wheels, sit on rail tracks lining the aircraft floor. If a taller man or woman is seated in front of a child, for example, the cabin crew will have the ability to move an occupant’s seat several inches back via smartphone or tablet, allowing for extra legroom.
“While passengers come in many sizes, children, adolescents, adults, men, women and with large height differentials within these categories, seat spacing in the main cabin of passenger aircraft is generally uniform except at exit rows,” the designers stated in their patent application, submitted in November.
“The one size fits all seating arrangement can cause discomfort for tall passengers, while a child or relatively small adult may be seated in an identical seat at the seat pitch, with more than ample leg room and in relative comfort.”
The legroom adjustable seat, however, leaves the final spatial arrangement to the discretion of crewmembers, not individual passengers.
“Even a relatively small incremental increase in seat spacing for the tall passengers can provide additional comfort with no loss of comfort to the much smaller passengers seated in front of the tall passengers,” B/E Aerospace said.

To get the most out of the Attribute Dependency Technique, follow these steps:
1. List internal/external variables.
2. Pair variables (using a 2 x 2 matrix)

  • Internal/internal
  • Internal/external

3. Create (or break) a dependency between the variables.
4. Visualize the resulting virtual product.
5. Identify potential user needs.
6. Modify the product to improve it.

The Attribute Dependency Technique: Three Ways to Create Smart Products

Published date: June 15, 2015 в 11:03 am

Written by:

Category: Uncategorized

Tags: ,,,,

When using the Attribute Dependency technique, you’ll reach a point in the function follows form process where it’s time to make adaptations to your concept. That’s where you try to improve the concept and put more definition around it.
One way to make adaptations with Attribute Dependency is to change the type of dependency. There are three ways to do it: passive, active and automatic. Think of these as what has to happen within the product or service for the dependency to take place. Let’s look at each type.
Mixing bowlsPassive dependencies, just as the name implies, are passive. Nothing has to happen for the dependency to take place. There doesn’t need to be an intervening element to cause the dependency.
Look around and you will see that many products and services are examples of passive dependency. Here is a simple example of mixing bowls that come in different sizes.
Now you may ask, “Is this really an example of the attribute dependency pattern?” It certainly is. As one thing changes another thing changes. In this case, as the needs of the user change, the size of the bowl changes. It’s a passive dependency, though, because the bowls simply exist in various sizes and shapes. In fact, any product that comes in different sizes such as clothing, hardware items, even homes are examples of passive attribute dependency.
Happy hourBut some dependencies require an active, intervening element to cause them to occur.  A very simple example is Happy Hour, when the price of drinks in a bar is reduced. But for this to happen, somebody has to do something. That active element, of course, is the bartender. At the appointed happy hour, let’s say 5 o’clock, the bartender simply lowers the price of the drinks presumably for an hour. Then again at 6 o’clock, the bartender raises those prices back to their normal level. Because of the active intervention, we call this an active dependency.
And finally, we have automatic dependencies. These are unique because they happen, as the name implies, automatically. The product or service is designed so that as one thing changes, the product automatically changes by itself without some intervening third-party element to make that change.
Transition sunglasses are one of the best examples of an automatic dependency. As the brightness of the light changes, the lens automatically darkens in response to that change.
Products that have this type of dependency seem almost smart. They know when it’s appropriate to change in response to some other variable, either an internal or external. The consumer doesn’t have to do anything because the product does it all by itself.
How do you know which type of dependency to use? It depends on a lot of factors such as how much convenience you want to deliver to the customer. Is it technically feasible to create a particular dependency? For example, your engineers might be able to make a mixing bowl that automatically expands as you put more things in it. But that also adds a lot of cost and complexity. It’s probably a lot easier for the customer just to grab the right size bowl to make a cake.
It also depends on how much control you may need in a situation. Do you want the customer or another person making the change? Look back at the happy hour example. You could create a cash register that automatically adjusts the price of drinks based on the time of day. The bartender wouldn’t have to think about. You would have complete control over the prices throughout the day.
Passive, active, and automatic. That’s three ways to give your customers very cool products with the Attribute Dependency technique.
 

Innovation Sighting: Attribute Dependency and World Population

Published date: February 16, 2015 в 11:17 am

Written by:

Category: Uncategorized

Tags: ,,,,,

What if countries were sized proportional to their population? What would the world look like? Take a look at this map (reported by NPR.org):
It’s a nice example of the Attribute Dependency Technique, one of five in the innovation method called Systematic Inventive Thinking (SIT). It’s a great tool to make products and services that are “smart.” They adjust and learn, then adapt their performance to suit the needs of the user. Attribute Dependency accounts for the majority of innovative products and services, according to research conducted by my co-author, Dr. Jacob Goldenberg.
Reddit user TeaDranks created this cartogram by creating a dependency between a country’s size and population. Each square represents 500,000 people.
To see more examples of Attribute Dependency with world maps, visit the website Worldmapper. It has “hundreds of cartograms, showing countries sized by everything from the number of books published or tractors working to condom use by men or woman.”
To get the most out of the Attribute Dependency Technique, follow these steps:
1. List internal/external variables.
2. Pair variables (using a 2 x 2 matrix)
Internal/internal
Internal/external
3. Create (or break) a dependency between the variables.
4. Visualize the resulting virtual product.
5. Identify potential user needs.
6. Modify the product to improve it.
 
 
 

Innovation Sighting: Attribute Dependency in Smart Apps

Published date: November 17, 2014 в 3:00 am

Written by:

Category: Uncategorized

Tags: ,,,,,

Are online reviews going extinct?
From Yelp to Amazon, reviews these days are good for just one thing: Seeing what others think of a product, service, or business. But are reviews really helpful?  Could they be an outmoded one-size-fits-all solution in a world where a user’s interests are increasingly customized and niche-specific?  Are they going the way of the dinosaurs?
According to ‘HeyLets‘ CEO Justin Parfitt — an expert on how to use reviews to make good consumer decisions — the next generation of review sites and apps will more intelligently utilize your personal data and contextual preferences to make more thoughtful recommendations.
It’s a perfect example of the Attribute Dependency Technique, one of five in the innovation method called Systematic Inventive Thinking (SIT). It’s a great tool to make products and services that are “smart.” They adjust and learn, then adapt their performance to suit the needs of the user. Attribute Dependency accounts for the majority of innovative products and services, according to research conducted by my co-author, Dr. Jacob Goldenberg.
HeyLets (www.heylets.com) helps you do the following:
1) Shows you a personalized feed of recommendations from users who have similar interests.
2) Uses your social data to inspire you to try new things across the full range of your interests.
Even more impressive, next-generation apps like HeyLets will soon learn over time how you live your life, and be able to do things like:

  • Anticipate your needs and propose activities for particular days by using information about past movements and even the weather forecast.
  • Automatically disregard reviews from people with distinctly different preferences (i.e. a vegan diner who posts a at a non-vegan restaurant).
  • Help you avoid less reliable reviews from “Debbie Downers” — people who only post critical updates and negative content.

To get the most out of the Attribute Dependency Technique, follow these steps:
1. List internal/external variables.
2. Pair variables (using a 2 x 2 matrix)

  • Internal/internal
  • Internal/external

3. Create (or break) a dependency between the variables.
4. Visualize the resulting virtual product.
5. Identify potential user needs.
6. Modify the product to improve it.

Creating New Opportunities in the Digital Space

The SIT method is great for creating exciting new products and services. But you can also apply these techniques to digital assets.

For example, let’s apply the Attribute Dependency technique to a website. You start by listing the internal and external attributes of the site. You list the attributes, and you create a two dimensional matrix that pairs internal attributes to other internal and external attributes.
Next, select a cell on the matrix and imagine a relationship between the two attributes. For example, “location of visitor” and “graphics,” meaning how the information is displayed on your website. As the location of the visitor changes, the information and graphics that you display on your website changes. Why would that be valuable? In what situations would it make sense to have that relationship in place?
Think about it. Imagine if your customer is browsing your website right inside one of your retail stores. Perhaps you would change the kind of information and graphics you would use to show your products.
What if they were browsing your website from one of your competitor’s stores? Could that change how you display competitive pricing information?
What if your customer is browsing within a healthcare facility, or from an airport, or inside a restaurant? Would it change the products, the prices, or other service elements that you display? It just might.
Applying attribute dependency can make your website responsive and adaptable. It services your clients better by understanding more about them.
Let’s apply this same approach to a social media application. For this example, let’s use Facebook. List the internal and external attributes of a Facebook Page, and create your matrix. Let’s imagine a relationship between “likes” and “wall postings.”  There is no relationship there now, so let’s imagine one. For example, as the number of “likes” increases over a particular period of time, your wall postings change. Why would that be beneficial?
Perhaps you would put different products or special promotions there once you reach a certain level of likes. In other words, you change how you engage with your customers who visit your Facebook page based on how they engage. A relationship between these two attributes would give you a cue to know when it’s appropriate to do something different on your page.
Let’s go further with digital innovation and look at mobile apps and how to apply SIT techniques.
For these, I like to use the Task Unification technique. In that case, we took a component of a product or service and we assigned it the additional job of addressing a specific business issue.
You can do the same thing with mobile apps. We create a virtual product by saying: the App has the additional job of addressing this business issue. The trick is to pick an app that has absolutely nothing to do with the issue now. That’s where you find some surprising innovations.
Let’s do an example. Imagine your company makes a household product that helps get rid of odors in your home. It’s a spray product that you would use to get rid of odors from your cat or dog. Imagine you’re the marketing manager for this product and you want to find creative ways to promote its benefits.
First, find a list of mobile apps. You can find many on iTunes, or a site like this one: Gotoweb20.net.  Pick one of these randomly and plug it into the phrase: the app has the additional job of promoting my product.
Here is an app called Micello. Micello is a provider of comprehensive indoor venue mapping. It’s like Google maps only for in indoor spaces like shopping malls or airports. You imagine this app has the additional job of promoting your spray for pet odors. What would be the benefit? How it would work, and how would it increase brand awareness of your product?
Suppose this technology is used to create an internal map of your home. What if it could also track where your pet spends its time as it moves from room to room. Perhaps the app creates an odor heat map of where the pet has been so that you know exactly where to spray the product.  I love this idea because it’s both functional and it reinforces the brand promise.
Task Unification can help find new uses for existing apps, and it can help you create completely new apps.
Your digital assets are just as important as your products and services. Using the SIT Method will unlock more value for your customers and find new ways to engage them more effectively  through digital channels.

Filtering Ideas to Find the Very Best Ones

The SIT Method is designed to help you generate lots of ideas in a systematic way. But how do you select which ideas to pursue? Filtering ideas is an essential part of the creativity process. You want to make sure you spend your time only on those with the most potential.

First, put all your ideas in a standard format. That’ll make it a lot easier to evaluate them. I like to use a template like this:

  • Name of Idea:
  • Description of Idea:
  • Benefits:
  • Target Audience:
  • Challenges:

Every idea should have its own name, not just a number. Give it a name that will help people see what the idea is about. Use literal names, not vague or confusing ones.

Next, put every idea into one of three categories. The first category is for those ideas that are a bit far out, perhaps borderline crazy. They’re novel, but they may not be feasible.

The second category is for those ideas that are just the opposite. They’re not wild at all. They’re incremental improvements.

The third category is for ideas in the middle – not too far out and not too near in. They’re in a special zone we call the sweet spot. They’re viable and creative. It’s these ideas that people get excited about.

But we’re not done yet. Once you put the ideas in these categories, look at ways to get more of them into the sweet spot. Here is what I suggest you do. Start with the far out ideas. Is there a way to pull them back in, take out some of the weirdness of the idea to make it more feasible? What if you eliminated an exotic feature of the idea but still retained the essence of what the concept is trying to do? That might eliminate some of the riskiness of the idea.

For those incremental ideas, find a way to push them out and add some novelty. For example, what if you used Task Unification to have one of the components doing an additional task? Or what if you applied Attribute Dependency to the concept to make it smart and adaptive? That would certainly add some novelty and push it closer to the sweet spot.

After this exercise, you’re ready to start evaluating your list of ideas. There are two ways to do this. One is very simple and informal. You ask a group of people to vote on the ideas. You have probably seen the so-called dot method. Here’s how it works.

First, let the group read the entire list of ideas with all the benefits and challenges. Then, each participant is given a number of small, sticky colored dots. They’re instructed to place these dots on the ideas they think are best. I usually have participants place these right onto the paper with the list of ideas. This keeps the voting anonymous and makes it more objective. Then, collect all the votes and tally them up. While it may sound overly simple, the dot method of voting has a lot of benefits. Each individual has their own biases of what makes a great idea, and they vote accordingly. But voting as a group tends to neutralize those individual biases. Many times, the group vote will tell you which ideas the company will prefer.

The other method is more formal and quantitative. First, create a scorecard by listing the four or five most important criteria for judging good ideas. Criteria might include how novel the idea is, how useful it is for your customer, how viable the idea is to implement, and perhaps how risky the idea is. For each criterion, use a rating scale of 1 to 4 where a 4 is highest and 1 is lowest. Don’t use odd number scales like 1 to 5 because people may have a tendency to overuse the middle of the scale and rate too many ideas a 3. You want to force their ratings to be on one side or the other.

Ask people to use the scorecard and rate each idea. Then, using a tool like Excel, put the data in a spreadsheet so you can calculate the averages of all raters. Add up the final score for each idea.

  • Novelty 4
  • Usefulness 3
  • Viability 3
  • Risk 2
  • Final Score 12

The ideas with the highest scores are your best ideas assuming you selected the right criteria. This approach takes more time, but it gives you more precision especially when evaluating a large pool of ideas.

True innovators generate great ideas, but they also use the wisdom of others to help evaluate them.

Innovation Sighting: Getting Your Competition to Promote You

Published date: February 24, 2014 в 3:00 am

Written by:

Category: Uncategorized

Tags: ,,,,,,

How do you get your competitor to promote your value proposition? By thinking inside the box, or, in this case, using the box.

DHL did just that in the highly competitive package delivery category. Shipping companies compete on the basis of speed, convenience, and reliability. So the race is on to prove to the market which company performs the best.

In this campaign, DHL spoofed its competitors like UPS to broadcast that it's faster. Can you guess how?

This clever campaign is an example of two of the five techniques of Systematic Inventive Thinking – Task Unification and Attribute Dependency.

The Task Unification Technique is defined as: the assignment of new tasks to an existing resource (i.e. any element of the product or its vicinity within the manufacturer’s control). In this example, the"competition" has been assigned the additional task of "promoting the DHL value proposition" about being faster.

To use Task Unification:
1. List all of the components, both internal and external, that are part of the Closed World of the product, service, or process.
2. Select a component from the list. Assign it an additional task. Consider ways to use each of the three Task Unification methods:

  • Choose an external component and use it to perform a task that the product already accomplishes
  • Choose an internal component and make it do something new or extra
  • Choose an internal component and make it do the function of an external component (effectively “stealing” the external component’s function)

3. If you decide that an idea is valuable, you move on to the next question: Is it feasible? Can you actually create this new product? Perform this new service? Why or why not? Is there any way to refine or adapt the idea to make it more viable?

The Attribute Dependency Technique is defined as: the creation/removal of symmetries or dependencies between existing product properties. As one thing changes, another thing changes. In this example, DHL created a dependency between "elapsed time" and the "visibility of the message."

To use Attribute Dependency, make two lists. The first is a list of internal attributes. The second is a list of external attributes – those factors that are not under your control, but that vary in the context of how the product or service is used. Then, create a matrix with the internal and external attributes on one axis, and the internal attributes only on the other axis. The matrix creates combinations of internal-to-internal and internal-to-external attributes that we will use to innovate.  Take these virtual combinations and envision them in two ways. If no dependency exists between the attributes, create one. If a dependency exists, break it.  Using Function Follows Form, try to envision what the benefit or potential value might be from the new (or broken) dependency between the two attributes.

Systematic Innovation at the Consumer Electronics Show

Published date: January 13, 2014 в 3:00 am

Written by:

Category: Uncategorized

Tags: ,,,,,

One way to develop your expertise in SIT techniques is with pattern spotting. A key premise of SIT is that for thousands of years, innovators have used patterns in their inventions, usually without even realizing it. Those patterns are now embedded into the products and services you see around you, almost like the DNA of a product. You want to develop your ability to see these patterns as a way to improve your use of them.

There’s probably no better place to practice pattern spotting than at the Consumer Electronics Show (CES). In last week’s CES in Las Vegas, “manufacturers demonstrated a range of previously mundane but now smart, web-connected products destined to become part of daily domestic existence, from kitchen appliances to baby monitors to sports equipment,” as reported in The Independent.

The word, “smart,” should tip you off right away. That’s a tell-tale for the Attribute Dependency Technique. It works by taking two attributes of a system and creating a correlation between them. As one thing changes, another thing changes. It tends to yield products that change or adapt to some changing need of the consumer. Hence, the product appears smart.

See if you can spot the Attribute Dependency Technique is these examples from CES:

  • Smart cars will become so smart they can drive themselves, avoiding congestion or collisions – even finding the closest parking space to your destination.
  • Smart refrigerators will let you know when the milk is on the turn, or when you need to buy more ketchup.
  • Smart toilets will monitor the frequency and consistency of your bowel movements, and tell you whether you ought to book an appointment with a dietitian – or worse, a clinician.
  • Smart ovens will manage mealtimes, cooking different dishes by different methods at the correct time.
  • Smart toothbrushes keep track of your brushing habits – not just the frequency of brushing, but also the technique. It then sends the dental data it has collected to your smartphone, with notes on how to brush better.
  • Smart “onesies” are not only a sleepsuit, but also a baby monitor. It tracks its infant wearer’s temperature, breathing rate, body position and activity level. It can even be paired with a bottle warmer, which starts heating milk when the Mimo senses the baby is about to wake up.
  • Smart tennis rackets record the power of each shot, the position of ball-on-racket, even the amount of spin. That data is then displayed on a smartphone or tablet, demonstrating the details of a player’s game and thus illuminating potential areas of improvement.
  • Smart beds track your heart rate, breathing, snoring, movements and surroundings, building a comprehensive picture of your sleep patterns which it then sends to your smartphone, offering suggestions for how to sleep better the following night.

With enough experience using SIT, you’ll use pattern spotting automatically. You will see some new product or service and instantly your mind will try to search which of the five techniques applies. When you get to that point, you have what we affectionately call the SIT “virus.” It means you are well on your way to mastering the method.

Inside the Box: Applying Attribute Dependency to Pinterest

Published date: September 9, 2013 в 9:39 am

Written by:

Category: Uncategorized

Tags: ,,,,

Pinterest has joined the elite group of social apps
along with Facebook, Twitter, LinkedIn, Youtube, and Google Plus.
Pinterest is a Virtual Pinboard that lets you organize and share
all the beautiful things you find on the web.”  How popular is it?  It is the fastest site ever to break through the 10
million unique visitor mark.  A report by Shareaholic claims, “Pinterest drives more referral traffic than Google+, LinkedIn and YouTube combined.”  As of May 2013, Pinterest was valued at $2.5 billion.

There are many creative ways to use PinterestNew apps are emerging around it much like what happened with Twitter.  But to maintain growth, Pinterest needs innovation.  Let’s apply Attribute Dependency, one of five techniques of Systematic Inventive Thinking, to Pinterest.

To
use Attribute Dependency, make two lists.  The first is a list of
internal attributes.  The second is a list of external attributes –
those factors that are not under your control, but that vary in the
context of how the product or service is used.  Then, create a matrix
with the internal and external attributes on one axis, and the internal
attributes only on the other axis.  The matrix creates combinations of
internal-to-internal and internal-to-external attributes that we will
use to innovate.  We take these virtual combinations and envision them
in two ways.  If no dependency exists between the attributes, we create
one.  If a dependency exists, we break it.  Using Function Follows Form,
we envision what the benefit or potential value might be from the new
(or broken) dependency between the two attributes.

The attributes of Pinterest are:

PinterestInternal Attributes:

  1. size of board (number of pins)
  2. size of the displayed board
  3. number of boards
  4. description of board
  5. subject of pins
  6. number of likes
  7. number of re-pins
  8. number of guest pinners
  9. who following

External Attributes:

  1. time
  2. followers
  3. boards trending
  4. links to other social networks

The new concepts are:

1.  Push To FriendsPinterest pushes a notification to Facebook friends or Twitter followers based on a keyword in the description of the Pin.  This is a bit like RSS feeds into a reader, but different in that the Pinterest board owner gets to decide what gets pushed to friends.  There are some existing links between Pinterest and the other social networks, but an approach like this could make it much stronger and more valuable.

2.  Pin RecommenderPinterest finds and recommends new Pins to you based on keywords in your Pin or Board description.  It is similar to the “You Might Also Like…” feature on many web applications.  A new app called SpinPicks does something similar, but it does not pull from the inventory of images in Pinterest.

3.  Board CloudThe Boards of a Pinner change size depending on Likes and Followers.  This is similar to a tag cloud where each word varies in size depending on how often it shows up on a website or document.  Tag clouds help the reader quickly understand which words are most prominent or popular.  Twitter has a similar feature called Trendsmap.  Given the highly visual nature of Pinterest, I would expect users to be able to turn features like this on or off in their settings to give a more personalized experience.

4.  Twitter TrenderThe boards displayed on the viewers main page vary depending on what is trending on Twitter.  Twitter has become the “eyes and ears” of the world, and hot topics trend all the time.  Pinterest would read these trends and match them to Boards for display on the front page, perhaps as defined by the viewer.

Follow Me on Pinterest

Get our innovation model that has worked for 1000+ companies.

    No thanks, not now.