How PLM can join semantic enterprise graph?

April 10, 2014

plm-and-enterprise-graph

Connectivity is a key these days and graphs are playing key role in the development of our connectivity. It doesn’t matter what to connect – people, information, devices. Graphs are fascinating things. Actually, I came to conclusion we live in the era of fast graph development. More and more things around us are getting “connected”.

It is almost two years since I first posted about Why PLM need to learn about Google Knowledge Graph. The story of Knowledge Graph is getting more power every day. GKG is growing. It represents “things” in the knowledge base describing lots of topics – music, books, media, films, locations, businesses and many others. Part of Google Knowledge Graph is fueled by Freebase – large collaborative database of structured data. Originally Freebase was developed by Metaweb and acquired by Google in 2010. It is still not completely clear how Google Knowledge Graph built. You can read some investigations here. Nevertheless, it is hard to undervalue the power of Knowledge Graph.

Another well known and publicly developed graph is Facebook social graph. Last year I posted – Why PLM should pay attention to Facebook Graph Search. Facebook graph represents structured information captured from Facebook accounts. It allows to run quite interesting and powerful queries (also known as Facebook Graph Search).

In my opinion, we are just in the beginning of future graph discovery and expanded information connectivity. It won’t stop in social networks and public web. I can see graphs will proliferate into enterprise and will create lots of valuable opportunities related to information connectivity and efficient query processing. Semanticweb.com article Let Enterprise Graph Tell You A Story speaks about enterprise as a set of Facebook pages. It explains how we can build a graph story of enterprise communication, collaboration, people activities, related data and other things. Here is my favorite passage from the article:

Wallace relies on Hadoop and graph database technology, with network data represented as a property graph. “Property graphs are utterly, totally extensible and flexible,” she said, and “the system gets smarter as you add more data into it.” The enterprise social network data generates triple sets (that John Smith created X Document that was downloaded by Jane Doe) that get pocketed into the graph, for example, as is metadata extracted from relational databases. A general set of algorithms can find a user within the graph and calculate his or her engagement level – activities, reactions, eminence and so on. “We now have a Big Data service with a set of APIs so people can query the enterprise graph,” she set, and then run analytics on those results that can drive applications.

I found this aspect of graph development very inspiring. To collect enterprise information into graph database and run a diverse set of queries can be an interesting thing. If I think about PLM as a technological and business approach, the value of graph connecting different part of information about product and activities located in different enterprise systems can be huge. These days, PLM vendors and manufacturing companies are using a diverse strategies to manage this information – centralized databases, master data management, enterprise search and others. Graph data approach can be an interesting option, which will make enterprise looks like a web we all know today.

What is my conclusion? The growing amount of information in enterprise organizations will change existing information approaches. It doesn’t mean all existing technologies will change overnight. However, new complementary techniques will be developed to discover and use information in a new ways. Graph is clearly going to play big role. PLM strategist, developers and managemers should take a note. Just my thoughts…

Best, Oleg

picture credit semanticweb.com


PLM Sales Cheat Sheet

January 20, 2014

plm-sales-cheat-sheet

I have to admit – I don’t have formal sales education. My childhood was mostly influenced by math and tech. Technology has a smell of precision and knowledge. At the same time, sales appears to be manipulative. I can try to blame Soviet Union regime, but it doesn’t matter now. I was wrong.

The understanding of how I was wrong came later in my life. I learned to sell my software projects, application and services. I still cannot say I’m good in sales. I must thank few people in my carrier that helped me a lot to understand the nature of sales. I’m still learning.

None of recommendations I put below came from formal books. Actually, I never had a chance to read formal sales books. Probably I should. List below comes from my experience and involvement in PLM projects in variety of roles – technical, implementation, advisory, strategy, competitive analysis and more. Following these rules helped me to achieve goals.

1- Make "enterprise executives" friends and friends across other divisions in a company you are selling to. Don’t be engineering/tech buddy only. The championship in engineering system department is important, but you need to get the whole picture of product problems and profit in a company.

2- Learn to say "No" to engineers and R&D managers. Very often, PLM has engineering roots in the organization. You can easy get spoiled by engineering ideas. With all respect to engineers, these ideas are not always on the top priority list for CIO and can be far from business goals.

3- Prepare to come at least 7 times to meet your prospect customer. Sometimes you will feel repeating yourself, but in PLM implementation, it is often part of learning and convincing themselves about their decision. Also, be ready to answer on the question "what is PLM and why is it needed if company already spent million dollars to implement ERP". You need to have a good answer…

4- You need to become a source of knowledge about other PLM implementations, best practices, failures and successes etc. Make yourself a bit "techie" – it will help you to build your credit. Very often your prospect customer doesn’t know what to do and would love to learn from what you did for other companies.

5- Learn how to shutdown implementations if things are going wrong. This one is tough. Sh’t happens. You need to learn how to fire customer, even if you made a sale. Not everything depends on you. PLM touches different departments and often requires a company to change the way they do business. People are trying to manipulate and it resulted in politics influence, conflicts between management groups, implementation strategies, competitors, etc. If you see that you cannot make a visible success in 3-4 months, push the stop button and ask to "rethink what needs to be done".

6- Learn few key typical PLM implementations failures – team disagreement about product development process and data ownership, CAD integration failure, PLM/ERP integration failure. In my view, these three are responsible for more than 50% of PLM implementation failures. You need to learn how to smell it and go to option #5 with request to rebuild the process.

7- Deliver one feature from engineers dream list. Take one that will help engineers to be proud of what PLM does. Something that not on the top priority list. You will become "engineering hero".

8- You biggest win will come 3 years later from a customer you failed to sale to. Enterprise sales is a lengthy process. Companies need time to understand how they do work and how they need to manage a change. Management changes. Corporate conflicts get resolved. Remember #4 and be consistent in PLM vision you sale. Prospects will come back.

What is my conclusion? Enterprise sales is a special discipline. These days is probably one of the most interesting things in the overall process of technological disruption. Major disruption happened because of internet, cloud, mobile and other technologies. The wave seems to be going to enterprise continent. However, enterprise is first about people and second about technologies. Technologies make sense only after people can understand and use them. Just my thoughts and good luck!

Best, Oleg


Who will make enterprise PLM UX more user friendly?

December 23, 2013

plm-user-experience

For long time, enterprise software didn’t pay much attention to user-friendliness. I can go and bring many examples from the past 10-20 years, but assume it is probably not necessarily. Consumerization of IT, mobile devices, internet made a shift in the way people started to think about future of user interaction, user experience and application design. Interaction with user before design phase, multiple user tests, re-design sessions – this is a normal way to design for a good interaction. You need to test your application with exact behavior of your users. I found the best example of what does it mean reading TechCrunch article on my way back to Boston few days ago – Cognitive Overhead. Interesting reading. The author is a founder of well-known app – Bump. The most amazing example for me was the fact they tested Bump on drunk people in San Fran and Palo Alto pubs. Here is the passage:

The very young and the very old are even more sensitive to cognitive overhead, as their brains aren’t accustomed to the sort of logical leaps our products sometimes require. Grandparents and children make great cognitive overhead detectors. When you can’t find old or young people, drunk people are a good approximation. In fact, while building Bump 3.0, we took teams of designers and engineers to bars in San Francisco and Palo Alto and watched people use Bump, tweaking the product to accommodate.

Enterprise software vendors these days are also thinking about the future of user friendly design and how it will impact enterprise apps. Information week article Infor bets on user friendly design to disrupt enterprise applications market speaks exactly about that:

Typically, people who decide on buying the enterprise application seldom use it. That’s why most people hate using enterprise applications. We want to get people using and accessing applications quickly, and that calls for a great user design. Social has huge relevance in this context. For example, a product called Infor Ming.le, allows employees to communicate, collaborate, and share information such as documents, plans, photos, and videos from a centralized location, with all activity captured and easily searchable.

It made me think about the problem of enterprise UX thinking and… learning. You need to learn from other people, learn trends, interact with other people working on user experience. So, what is the future of education in design and user experience? When I was looking for the answer, another article caught my attention related to the future of UX – The Top UX prediction for 2014. Interesting enough I found some answers on the future of UX design training and education. Here is my favorite passage:

Meanwhile, product design is now commonly recognized as a strategic advantage, its business impact made obvious to even the most skeptical of analysts by the success of Apple. Ironically, as companies have become more design savvy, some designers have felt marginalized when early stage, strategic product design decisions fall to business executives and product management. We are going to see more and more colleges recognizing the importance of customer service and experience in academia. A business degree, even a masters or PhD, with a specialty in customer experience design will become a new offering.

What is my conclusion? Application design is getting wide recognition among companies in all domains – consumer, enterprise, education, training. I can predict lack of educated people that can do it professionally as well as high interest in training sources focuses on product design. It will be specifically hard in such complicated domains as enterprise applications and PLM. So, who is going to be the next "Jony Ive of enterprise"? This is a good question to ask. Just my thoughts…

Best, Oleg


The Ugly Truth of Multi-BOM Management

December 18, 2013

multiple-bom-ugly-truth

Bill of Material (BOM) management is always fascinating topic. It sparks so many debates and introduce a large set of diverse opinions. I can even say that I have a special passion to speak about BOM on my blog. If you want to catch up on my recent posts about BOM, you can try these few links – Will PLM manage enterprise BOM? and Will SaaS and Open API solve BOM management problems? My special passion is "single BOM". I started this conversation few years ago. Here is my last writeup about single BOM- Single BOM in 6 steps.

Few days ago, my attention was caught by PLM dojo article about pros and cons is Multiple BOM management – Why You Should (or Shouldn’t) use Multiple BOMs. I highly recommend you to have a read the article including comments (the number is growing). It brings an interesting set of strategies relasted to BOM management. From my side, I can clearly see advantages of both approaches. And I can generally say it depends on many factors – industry, product, organization, processes and… (what is not less important) people. Here is my favorite passage:

Sometimes it makes sense for the CAD user to organize the design differently than how ERP organizes the data. For example, it might make sense to group a large assembly model into sub-assemblies that don’t represent any actual part, but make it easier to divide up work on the overall structure. A related reason is that having the part BOM separate from the CAD BOM isolates the part BOM from the inevitable messiness of the CAD files.

While there is nothing wrong in division and separation of CAD design and Part structures, I still believe there is a trick here. Thinking about that, took me back to the post I wrote few years ago – The Ugly Truth About PLM-ERP Monkey Volleyball The controlling of data is one of the fundamental enterprise software behavior and strategy. One of the "negative" aspects of single BOM strategies is the need (and complexity) to share responsibilities and control over the shared "single BOM". It can create lots of organizational constraints, especially if departments and/or divisions are using multiple systems.

At the same time, Single BOM containing multiple dimensions of product information can become a place to share data among organization and optimize processes. However, in order to make it happen organization will have to agree how to manage "shared space", and shared responsibilities. People management becomes a critical function to make it successful.

What is my conclusion? Technology is easy part, but people are really hard. This is one of my favorite quotes. The ugly truth of BOM management is the fact it requires people management and agreement across organization. Multiple BOM can be done using separation and data island controlling. Very often you can hear about technological challenges of single BOM organization. Much rare situation is when organization is moving to people and organizational constraints. People’s ego and organizational issues are often playing a key role in decision to go with one of BOM management strategies. Just my thoughts…

Best, Oleg


Full Product Lifecycle, Cloud and Granular Security

November 27, 2013

granular-security-cloud

Security is one of the most fundamental things in enterprise systems. If you think about every enterprise system as a big data store with applied business rules, security will be the one you will need to evaluate first in order to map it to your organization. It comes in multiple dimensions about who is getting access to information and in which way. Additional business complexity can increase the complexity of security model even more.

Cloud systems development is trending these days. We can see additional cloud applications and systems almost every day. It comes to the situation when collaborative usage of multiple cloud systems and applications becomes real. I’ve been reading Engineering.com article Can CAD-in-the-Cloud Handle the Entire Product Lifecycle? Read the article and draw your opinion. The important intent I captured is related to combining different disciplines and cloud tools under a single hood of entire product lifecycle development and business practice.

The decomposition of the process into stages and using of different cloud application can bring an interesting perspective on required security model. Actually, I can see security becomes a hot issue in the cloud development. The time when security was only about "file-share" options is gone. Today, cloud systems and users are demanding much more granular data organization and security management.

My attention was caught by an interesting acquisition made by Box. Navigate to the following article to read – Box Acquires dLoop To Enhance Security With Fine-Grained Data Analytics Technology. Box is well known outfit producing variety of tools and specifically focusing on enterprise customers. My hunch, the deal is the answer to an increased demand for data security in enterprise and it comes as part of growing competition with tools such as Dropbox and other consumer file sharing tools. Despite huge success in consumer space, security and enterprise deployment is probably still a challenging point for most of them. Here is an interesting passage from TechCrunch article:

We’ve been spending a lot of time improving the end user experience on Box, but we’re equally committed to creating better management tools for enterprise IT. This means unlocking greater visibility into the activities happening around information, and providing more granular controls where necessary. dLoop’s machine learning capabilities will ultimately allow Box to help enterprise customers identify and surface relevant content by tracking activity patterns… In larger enterprises, data classification is becoming a must-have in order to control what people can do with files. Companies want policy-based file sharing. Box has made considerable effort to enhance its security.

Policy based security is a key thing here. Thinking about full product lifecycle scenario involving multiple tools and people, policy based security can be the only way to support a granular security model. The needs for security is only one aspect. Ultimately, the goal of enterprise systems today is to improve user experience. In my opinion, without an appropriate data granularity, this is mission impossible for most of systems today. Cloud is adding an additional dimension of complexity. It comes as part of scenarios related to multiple cloud tools and shared content.

What is my conclusion? New environments, old problems and even bigger challenges. This is how I can see a combination of enterprise reality, new cloud systems and demand for security. The granularity is a key, in my view. Without policy based granular access, cloud product development tools such as CAD, CAE, CAM, PLM and others will remain childish and fail meet real enterprise customers scenarios. Just my thoughts…

Best, Oleg


The End of Single PLM and Enterprise Complexity

October 11, 2013

It is not unusual to hear about PLM as a single system capable to drive overall product development processes in the organization. Even more, if you look over PLM vendors marketing materials you can see many nice pictures and charts that will show you how to organize single PLM system. The reality is far from that point. Especially if you come to large enterprise organization. Enterprise is complex, driven by lots of requirements, enormously regulated and has to maintain connection with thousands of suppliers and people working in outside of the office facilities.

plm-marketing-charts

Engineering.com interview with Airbus’s top-ranking IT executives, Anders Romare brings lots of lights on how such a comlpex enterprise is organizing their product development and manufacturing activities as well as what PLM word means for Airbus. Navigate to the following article – PLM Systems Cleared for Take Off at Airbus, read more and draw your opinion. Airbus is using three main systems to accomplish their PLM goals. Here is a short passage that explains that:

Q: What are the major software solutions in use at Airbus? A:Catia for MCAD and digital mock-ups, with Dassault’s Enovia VPM as vault is the foundation, says Romare, while we manage and control product data via Windchill/PDM Link from PTC. Manufacturing data is managed by SAP on the ERP side, in conjunction with PDM Link. Then there are of course a number of other applications for specific needs, such as simulation. But the overall foundation is based on software from SAP, PTC and DS.

Modern enterprise is diversified. PLM environment must be able to provide an access to thousands of organizations outside of firewall. It brings the complexity of security, data access, sharing and computation to extremely high level. The interesting number for was to learn that 85% of PLM system users are external. Here is another interesting passage:

We must, for example, ensure that approximately 20 000 users are able to reach and access the PDM data. Not only that, we can have as many as 4 000-5 000 concurrent users, in real time, that share 3D data in context. This in turn requires a lot of computing power, especially when you consider that important parts of our product development work is focused on simulation, something that we enable with one of the world’s largest corporate-driven HPC facilities (High Performance Computing).

There is a single keyword explaining everything in the enterprise organization – complexity. If you think about data processing, you see a sequence of at least 3 systems processing CAD structures and Bill of Materials – Enovia VPM, Windchill and SAP.

In the 3D assemblies of the design stage, Airbus treats the engine as a "shell" to which a number of access points for mechanical, electrical and software is to be added. The responsibility for motor the area is held by the engine supplier and the aircraft operator. On the MCAD side DS ENOVIA VPM serves as vault and collaboration platform on mock-up level, but reporting under the umbrella of PTC’s Windchill/PDM Link. All of the general process and data management tasks and things like configuration management are managed directly in Windchill. Even the bills of materials (mBOM and EBOM) are planned and configured in Windchill, for instance when downloading Catia product data, although these will not be accepted until they’ve passed through the SAP ERP system. Note that Airbus also uses a range of software other than the above.

The article about Airbus PLM implementation made me think about two fundamental challenges of enterprise PLM software today. The first is related to the lifespan of the system. While most of software vendors are release software on annual basis, enterprise customers are not interested and not capable to use it. Release of software is expensive process. With no much between release and consumption, significant amount of software development budget is going to be wasted. The second one is related to desired level of complexity in data management – 1000s of users, highly complex data, requirement to provide a global access to highly configured data. All these things together make data management problem bigger than ever.

What is my conclusion? In my view, enterprise software development practices should be reviewed. There is no sense to release a product while customers won’t be able to use it. Maybe an alternative is to move from annual software release into continuous release practiced by many cloud vendors? Data management complexity of enterprise organization can compete well with challenges of global web brands. It is a time to look on the web as a source of future innovation in enterprise. Just my thoughts…

Best, Oleg


Why you still cannot sell PLM without exec support?

August 6, 2013

There are variety definitions of what is PLM. Not much agreement about that among PLM vendors, PLM analysts and manufacturing companies. However, if you want to introduce PLM in manufacturing company, you can find majority of people to agree that implementing PLM takes time and includes some strategy planning, implementation and patience. My PLM industry colleague and blogging buddy Jos Voskuil calls it PLM journey. One of the things that often comes during PLM journeys is to have C-level support during PLM sales and implementation.

Aras Corp. recently featured the issue of C-level management support for PLM in their blog – Executive Support for PLM and Why You Need It. I captured few interesting quotes from that post:

Executive support is a crucial part of the PLM selection, implementation and acceptance processes that often gets overlooked. Having executives back a PLM initiative not only shows that the software is here to stay, it also ensures resources, and brings executive insight into the business model and processes to the project.

PLM is not a one-year solution; it’s a long term decision and the results have to be measured over time. For example, are your product development cycles still improving 3 years later? ROI has no expiration date. Including the long term impact of PLM in your business case could be the key to getting much-needed executive support.

In my view, these messages are capturing very well the essence of what called "PLM business transformation" process. You cannot sell and implement PLM without having PLM implementation champion in a company as well as execs giving their buy-ins to the project. It represents what I called enterprise sales old schoolers in one of my previous blog post. However, I’d like to put some controversy in this PLM journey nirvana. It would be great for business to have time to choose their PLM journey road with enough time for discussing, planning and implementing. It should be also good to have unlimited resources to spend in PLM journey. However, this is probably not going to happen these days and for visible future for most of companies I know. Businesses are very dynamic these days and driven by lots of unpredictable business factors, cost constraints, competition and regulation. It applies a new perspective on how companies will treat PLM in the future. And, it will drive even more attention from executives in the companies.

What is my conclusion? Regardless on what PLM implementation strategy you choose these days, to have exec support for your PLM project is a good thing. When it comes to redefining how is you company develop and manufacturing products, you want to have some C-level support people to backup your PLM strategy. In the modern dynamic business environment to have execs on board of your PLM implementation is even more important. Nowadays, you don’t have 5 years to manage your PLM journey project. Your PLM journey becomes PLM sprints with execs serves as supporters and judges at the same time. Just my thoughts…

Best, Oleg


PLM and Google enterprise numeric factoids

July 25, 2013

Google is making lots of things these days. The list includes search, data centers, mobile phones, tablets, wearable devices, self driving cars and more… The question about Google and PLM is one that very often drives rumors and lots of speculations. It usually comes from the side of analyzes applied on Google enterprise business. Here is the next round of discussions. I captured this story in ArnoldIT blog – Google and Its Enterprise Push: Changes Ahead? The assessment is that Google is probably making around $1B of their revenues driving variety of enterprise programs.

Google’s enterprise business consists of productivity apps like Google Docs and Gmail and the Google Cloud Platform, which includes its App Engine platform-as-a-service and its Compute Engine infrastructure-as-a-service. One important factoid in the write up was the assertion that Google’s cloud business and its applications business was generating about $200 million in the most recent quarter. Assuming that the figure is accurate, Google is on track to generate about $1 billion from its enterprise services. If the company ends the current fiscal year in the $60 billion in revenue range, the enterprise unit will make up one minute of 60 minutes of Google revenue.

The interesting numeric comparison is that Google’s $1B of enterprise revenues is compatible with 4 top CAD/PLM companies ($1-2B). At the same time, increased investments in hardware, wearable computers, cars and other physical devices can make Google enterprise execs to pay attention on what companies in CAD/PLM area are doing.

What is my conclusion? Google was puzzling around enterprise business already few years pushing and pulling services and offering, changing prices and strategies. Are we going to see next round of Google push in the enterprise? Will Google try to leverage Microsoft management crisis to capture few additional points of enterprise apps selling. Maybe Google will decide to invest money in CAD/PLM business to make their hardware programs more efficient? Who knows… These are just rumors and my thoughts…

Best, Oleg


Why PLM needs to shift focus from buyers to users?

June 3, 2013

Enterprise systems for long time are well-known as a place where IT plays the role of the king on the road. ERP, CRM and many other enterprise systems and implementations proved that. If you want to sell to enterprise organization, you need to focus on key IT people, preferable on CIO, Chief of Engineering, Manufacturing etc. Earlier this year, I had a healthy debate about this topic following my blog post – PLM, Viral Sales and Enterprise Old Schoolers.

The disruption in Enterprise IT is one that I see as one of the most interesting trends these days in enterprise space. The following presentation caught my attention yesterday – The challenges and opportunities of business in the disruptive tech era. I recommend you to take a look on this even the presentation is 56 slides. However, the following slide stands out and resonate with the point I wanted to make about IT.

Let’s get back to PLM domain. In the existing ecosystems, there are two major ways to sell and implement PDM/PLM projects. One can be made indirectly mostly via CAD vendors channels. The complexity of these implementations is limited and these implementations (with some small number of exclusions) are limited to catch the level of enterprise IT. Another one is a direct channel developed by PLM and ERP vendors selling PLM implementations to top level management in IT organizations. The higher level of IT people is better.

I can see multiple reasons why existing IT is not getting excited about technological disruption in PLM and other enterprise organization. The disruption means changes and changes are usually come with the lost of control and existing status. For example, cloud means no servers need to installed, implementations can be done remotely and product development has a better chances to focus on user experience and business needs rather than on how to implement and run enterprise deployments.

What is my conclusion? The future of PLM implementation will shift focus from PLM buyers to PLM users. At the end of the days, people need to get job done. PLM needs to focus on user needs, user experience and the ability of systems to help people in everyday business life. Just my thoughts…

Best, Oleg


Legacy Systems and Future Coolness of PLM Software.

May 30, 2013

Enterprise software sucks. How many times we’ve heard that for the last 5 years? Probably too many… I remember one of the first articles about that back in 2007 – Why Enterprise Software Sucks? by Jason Fried. The article got almost 100 comments, which confirmed that the topic does matter. Earlier this month, I found an article re-thinking the view on enterprise software – Why Enterprise Software sucks – 6 years later. Take some time and read these articles. There is a clear confirmation that trend towards developing better enterprise software is positive. While you can still see lots of crappy software used by enterprise organizations, enterprise software and SaaS is seeing lots of innovation. At the same time, legacy systems is one of the main factors that prevents enterprise software from innovation. Here is a very important passage:

One of the main reasons that enterprise software sucks is because enterprises are still using software from the 1990′s (or worse). Putting traditional enterprise systems in place is expensive, time consuming, and requires massive customization. Once it’s deployed, it gets relied on by other software, workflows, and processes. Over time, it becomes mission critical and a load bearing wall. Companies are entrenched in their systems and don’t dare touch it if it’s “working” (think: Windows XP).

Let’s get back to PLM. Think about typical situation of manufacturing company using PDM/PLM legacy systems. The deployment cycle of legacy PDM/PLM systems is 3-5 years from initial product introduction and demonstration. Installation and customization can potentially take up to 1-2 years to get it done. As a result of that, companies are using PDM software developed back in early 2000s. The average investment into an existing PDM/PLM deployment can be in the range from $0.5M up to millions of dollars. The level of complexity to update an existing system is high. The migration is combined from multiple factors – data import, customization of new systems, people training. Companies need to pay a lot of money to re-implement software and processes, which is a complicated decision in terms of product ROI.

What is my conclusion? The roots of future enterprise software coolness is deep in enterprise legacy products. The ability to transfer data and processes from old systems to new environment and keep ROI of these projects high – this is a major goal of new enterprise vendors these days. Just my thoughts…

Best, Oleg


Follow

Get every new post delivered to your Inbox.

Join 218 other followers