Presenting sensitivities

I added a few more slides to the database today, one of them was this one: a template to present a sensitivity analysis.

Why this particular layout?

• It is nicely spaced out, a calm composition for so much data

• Numbers are disconnected from the spreadsheet: rounded up, entered by hand

• Colours, bold, are used to direct the eye to what is important, and what is secondary

Some thoughts about how and when to present this type of analysis:

• Presenting sensitivities and not the same as analysing them. The latter is the homework that you should have been doing before the presentation. Figure to what factors your model is sensitive, decided whether that is how it should be, then gather more information where needed to increase your confidence in variables that can make a big difference. What is left to discuss are sensitivities after you did everything to minimise and/or understand them.

• The ranges of the variables you show should be realistic. This is not an exercise in mathematics, but an attempt to really understand what drives the future.

• Pick dimensions that are not correlated, if the risks on the x and y axes are the same, you are not adding much insight.

• Try flipping the analysis upside down, instead of showing deviations from the base case, show “what you would have to believe” in order to get to a certain number.

• Be careful when sharing this type of data if you are in some negotiation about valuation. If the other side understand your model, they can basically salami slice the valuation using your own excel. You need to understand the sensitivities, but sharing them directly might not be smart.

The slide layout is available inside the SlideMagic app, or here on the web site.

What it took

Building a product is slow, but steady going. I jotted down this list of the various hurdles and went through to get a useable product today:

• Design the UI: most of this was done for version 1.0 five years ago

• Understand the basics of Javascript (with 1990s Pascal to start from)

• Understand post 1990s programming concepts: objects, methods

• Get an environment up and running so that I actually could run a simple piece of code

• Find a way to get access to the data (presentations) version 1.0 was producing.

• Setup an environment that turns a program that says “Hello world” and turns it into a desktop app

• Figure out a way to scale text in a browser environment, preserving the exact proportions of design elements (resize your web page, this is not what most web pages need).

• Get github and multiple versions to work

• Build the first rendering engine that actually displays a chart: text boxes are easy, scaling images a bit trickier, data charts get nastier even

• Find a way to register clicks and make things editable: shapes, menus, in different context.

• Copy the rendering engine to a generic format (for thumb nails in story mode for example)

• Duplicate the app engine to enable multi-screen presenter mode (running 2 processes and a master process that talk to each other)

• Enabling on-screen editing of text, graphs, image dragging, image cropping, flipping

• Building the grid editing system (implementing my patent)

• Build the PowerPoint conversion

• Build the PDF conversion

• Build the image export

• Build the printing functionality

• Add automatic 16:9 to 4:3 and back conversion

• Add automatic dark/white background conversion (beyond simply changing the background color)

• Enable multiple windows (each window is a full copy of the render process) and coordinate settings between them

• Build user authentication: pro users get features others don’t have access to via a web server, involving password hashing and building a user database

• Build the first version of the online template database: search slide layouts inside the app, but pull data from the central server

• Hook up unsplash image search

• Hook up noun project icon search

• Create an auto-update mechanism that updates the desktop app in the background when new versions are released

• Get the mac app to run on windows as well

• Get certified with Microsoft and Apple so that people don’t receive scary warnings when installing the software

• Build the full-scale slide template server, integrating the PowerPoint-only content from the old one

• Get payments working

• Build the front end of the marketing web site and the template store

• Get PowerPoint conversion to work on the server as well

• Build the management console to manage slides, users, and the search algorithms

We are going into the next phase now with a live product, direct user feedback is exciting, but it also means I cannot pull off some of the dramatic overnight feature and architecture changes I did over the past year. To be continued.

Photo by Thomas Galler on Unsplash

The genie is out of the bottle

I just took the password wall off the SlideMagic 2.0 app, so everyone can have a look (www.slidemagic.com., you just need to create an account. Payment plans are working on the template store, which should also be reflected in the app. I am sure there are hidden bugs lurking, but I keep on pushing. Remember that I do not have a 24/7 support organisation up and running yet… Valentine’s Day 2020 is a memorable day already :-)

Photo by Cesira Alvarado on Unsplash

Accounting...

A quick update (things are a bit quiet on the blog this week).

While the SlideMagic app is my project that I can afford to take some time for to go through multiple beta iterations to get it right, the template bank is a much more straightforward business, I want to get it running on the new platform as soon as possible.

The key functions work: search a slide (search algorithm is still being improved), and you can download it in multiple formats. The “only” thing left to do is the payment infrastructure. I got the account management and download book keeping working which leaves me with one nice challenge: EU VAT accounting for digital downloads.

The EU put all kind of laws in place to get a cut of the profits that the internet giants are making over their head. That is all great, but in the process they created complicated accounting rules that also apply to me. I as a tiny startup need to have the same tracking in place as Spotify.

This is a technical challenge that will not move humanity forward (unlike the UI of my app, or the new slide search algorithms). But, I have to crack it. Hopefully in a week or so, my system accurately adds taxes (or not) depending on your country, as an individual, as a company, with valid (or invalid) EU VAT numbers, and generate reports for my own book keeping and refunding EU tax authorities at the press of a button. My accountant (a big international firm suggested I do this manually, I think there must be a smarter way to do this. Pfff.

Work in progress

Photo by Ray Reyes on Unsplash

How to add captions in PowerPoint, as you speak

I did not realise PowerPoint could do this: put captions or subtitles on your slide live, while you speak into a microphone. Instructions how to get this work can be found here.

Photo by Immo Wegmann on Unsplash

404 clean up

After putting the new site on my own server in the root of slidemagic.com, replacing squarespace, I got to see the amount of traffic my 10-year-plus blog is actually getting. First, I thought that I was under attack, but they all seem legitimate Google searches that keep on coming and coming.

Over the years, I have moved my blog across various platforms, and now I could see that the link conversion did not work very well, a huge amount of 404 / page not found errors. I let the server log run for 24 hours and by now have probably caught most errors with smart redirects. Let me know if you still encounter problems.

The logs also shows the ridiculous amount of automated hacking attempts. My site is not very high-profile (yet), but various IP addresses from the usual countries are constantly trying to find the Wordpress log in page (I am not using Wordpress) and other security leaks. Let’s I hope I did not leave a door open anywhere.

I am putting more protection in place at the moment, which might in turn result in short hick ups in the web site’s availability. Apologies for that, but hopefully it will prevent bigger problems in the future.

Business model for SlideMagic

I put up my first thoughts of the pricing and business model for SlideMagic on the beta web site. The basic logic:

• A free tier that allows for a basic desktop app install and access to the free templates

• A pro tier with a yearly subscription to the desktop app including pro features (PowerPoint/PDF conversion, in-app template search), and access to the full template bank

• A day pass with 10 pro slide template layouts (online only, not from within the app)

• The beta tier (to which some of you have registered), 30 day access to the fully fledge app, without access to the pro template bank

The table below summarises things:

I am pondering how to treat independent consultants that sell decks with SlideMagic templates in them to their clients. I think I will follow the model that most paying stock photo sites use: a new paying client mains a new payed license. So a consultant can use her license to make slides for her own firm (pitching projects), but as soon as she uses SlideMagic on a paying project, she needs to expense a new SlideMagic license to the client.

What do you think? Fair, practical?

Working like crazy without being tired

Back at McKinsey I had a grueling lifestyle: long hours and always feeling tired. Many years later, i think I put in the same amount of hours getting SlideMagic 2.0 up and running, but…. I am not feeling exhausted all the time.

The secret? Focus. No commute, no travel to (unnecessary) meetings, I can pretty much get 10-13 hours of productive time each day (6 on weekends).

Also it helps to focus on what you liking doing next: crack that difficult algorithm, make the site look better, do the accounting, it is all a different mindset that is not available on command. Being able to choose what to work on is great.

Work in progress.

Photo by Vladislav Muslakov on Unsplash

New site now live

The migration of the site to the root url slidemagic.com worked. I took the opportunity to clean out a lot of old links and analytics scripts that were still hiding in the squarespace template. The only thing that is still running on squarespace is this blog and the HTML version of my book from a few years ago. Everything is now my own server.

I am thinking very hard about moving the last bits over from squarespace, the site has become slow and buggy over the past year. Maybe it is better to host it myself, with a more minimal format, without the big banner images and glitter that we do not need. The challenge is to migrate 10 years worth of posts across… We will see.

Rebranding and juggling URLs

I am throwing around URLs over the weekend as I slowly position the new sever into place. Things might break, formatting might be strange. Work in progress…

Photo by sol on Unsplash

SlideMagic explained, well, in a slide

The diagram below is in typical SlideMagic style. While a professional designer can do better, the slide looks professional, everything lines up, has a clearly recognisable branding, and takes very little time to make.

I scribbled down this slide for myself, to see how to position the online template bank and the desktop app. This diagram is probably not the best way to start marketing SlideMagic, but it shows what I try to achieve for an “internal audience”. (And frequent blog readers probably fall in that category :-))

It took exactly 5 seconds to make the slide available in the template store and the app itself (you can find it here. I put it in the free tier for now, it is a good test case for the PowerPoint conversion quality, and the slides that the algorithm thinks that are related (still work in progress).

Lighter variants of your presentation color

I ran into this slide (fragment) presented on an online course site the other day (I now digest tons of these to refresh my coding skills):

It shows a common problem in PowerPoint: you picked a nice theme colour (deep purple in this case) and you need variants of it. (This presenter figured out that too many colours makes your slide deck cluttered, hence SlideMagic only allows one :-) ).

The default model to make colour changes is to modify its brightness. It almost always work to make things darker, the other way around though can create a problem for very saturated colours. You don’t notice the saturation level at dark levels, but on brighter variants, that elegant purple becomes cute/bright pink.

The solution: change colour saturation as well as brightness. This post on my blog from 10 years ago (what?) describes it:

PowerPoint 2010 gives you the option of a spectrum of different shades of the same color. This is great to design charts with a consistent color scheme.

However, if your template contains colors that are highly saturated, the suggested lighter shades of your color will be too bright to use as neutral color nuances. Here is how you can fix it. (Click on the image for a larger picture.).

Create a new base color by reducing the saturation (in laymen's speak: make it more grey). Open the color in your color template (format shape/fill/solid fill/color/more colors)Switch the color model from RGB (red, green, blue) to HSL (hue, saturation, luminance).Reduce the (S)aturation value, while keeping all other variables the same.Use a lighter shade of this new base color instead and save this as a new color in your color template.

A (by now blurry) image with additional clarification:

A simpler approach:iInstead of changing the RGB values of your colour, simply add transparency to the dark colour. This works great on white backgrounds, but will create problems on coloured surfaces.

Another approach building on this:

• Create a shape with the dark colour over a white background

• Increase the transparency until you reach the desired level of “brightness”

• Hover over the box with a colour picker and add the colour you just picked to your slide colour palette.

Photo by Denise Johnson on Unsplash

Freelancing and career stability

Now and then I get people considering a career change asking me about life as a freelance consultant. “Isn’t that a very uncertain profession, when compared to someone in a permanent position?”

My answer: not really. Yes, you need to get your first clients/projects, but if you do good work, clients will refer you to other clients, and clients will come back. A person in a permanent position has one employer (who can go through reorganisations and other fun things you do not control), whereas you can hedge and spread the risk among multiple “employers”.

There are a number of things to consider though:

• The biggest one: do great work. As small independent business, you depend on word of mouth advertising. A beautiful website, massive SEO efforts, might get you the first inquiry, but then people want recommendations. You simply write off an unfortunate purchase of a $5 book that was not great, a 1-month consulting project is different. Most internet marketing has very low funnel rates, freelancers rely on 95% conversion.

• While the downside for a good freelancer is limited, the upside is probably limited as well. It is hard to scale a bespoke service business. You want to add a second person and not dilute your offering, that person probably needs to be as good as you, and as a result will need to be paid as well as you. Two times the revenue, two times the cost, the same profit. For my own presentation design business, I realised 1 employee (myself) was a great business model, and 50 probably is. But the role of making sure that 49 people have enough work and do it well, is a very different one from helping clients yourself. In short, as a freelancer, you are likely to pretty much the same thing in a few years from now as yo do today, albeit better and a bit faster. In a big corporation, you can go up the ladder and do different things every couple of years.

• There are things that freelancers often overlook in their pricing: pensions, disability insurance, etc. etc. All this protection that a regular employment position usually offers.

In short, it depends…

Photo by Max Ostrozhinskiy on Unsplash

Working on slide searching and tagging

Some of the world’s major artists got popular because they documented their life and motivations for their work in letters (Van Gogh etc.). Maybe the same will happen with this blog, and my journey from helping 1 client at a time to get better presentations, to hopefully thousands/millions eventually.

So finally, finally, I got a tool in my hands that I wish I had 5 years ago already:

• Create slide layouts and variations super quickly

• All in a consistent format

• Some way to store and organise all the original files, PowerPoint conversions and slide image thumbs (in different formats)

• A way to tag, describe, and search slides, and also organise, slide, and dice this “meta data”

This was the “easy” part, but a required step to get on to the next phase. How do you actually best describe slides, what slides are related, what suggestions for other slides should you give a user.

My current setup is one big playground. The number of templates is now double that of the current paid Shopify store. I am changing things and experimenting with instant feedback (whether searches are relevant, how much time it takes to complete them.

For most content on the web, and documents in corporate databases, you can use a Google approach to slides: you get a dozen links or video thumbs, and a human eye can quickly go through the clutter and find that document she is looking for. For slide layouts, this is not good enough.

Work in progress.

Against the light

In the early 1990s at McKinsey, presentation design was actually document production. Hand-written sheets of paper would be entered into a computer by full time graphics designers. Each word, each line, each graph. Then the whole thing would be printed and bound in books.

I remember the final quality check of the Amsterdam office manager: holding the pages against a strong light to see whether the titles, footers, page numbers, and margins of the slides lined up. You were in trouble if they didn’t.

Getting these basics right is very hard in today’s PowerPoint, If you copy and paste slides between masters, the alignment of objects will be off. If you change screen sizes (from narrow to wide screen and back), things go all over the place. Or, if you use/buy other people’s templates, they won’t fit well in your company’s slide layout. This is not PowerPoint’s fault, any software that needs to give total design freedom to its users will have this side effect.

I went through this the hard way myself, as I am making the slides of my “old” template store compatible with the new format of SlideMagic 2.0. Hundreds of slides that require small corrections to get things to line up properly.

With SlideMagic, professional designers might complain about the lack of flexibility in layouts, the rest of us will be extremely happy with how easy it is to tweak templates, screen sizes, and copy slides between presentations.

Photo by Bank Phrom on Unsplash

It makes sense, but it does not

I have been in many of these types of presentations:

Sometimes a big company does something, and you know exactly what every slide in the PowerPoint looked like, and all the slides made *sense*, and followed on logically to the conclusion... except that somehow the conclusion makes no sense.

— Benedict Evans (@benedictevans) January 17, 2020

Some of the reasons why the overall conclusion of a presentation does not make sense, while the individual slides do:

• Maybe the people, the organisation, and its culture is not the right environment to make a plan happen. Who is going to do it?

• The probability curve: on average, normally speaking, the strategy makes sense. But what if things start deviating from the average. What is the potential downside and could it be catastrophic to the compony?

• The self-fulfilling prophecy. Yes, the deck has 50 slides, discussing 50 different aspects of the idea, but when you look at it, they all depend on ver few (maybe even one) assumption about the market outlook, a valuation of a company, etc. That assumption could be wrong.

Unlike for big companies, for tiny startups the opposite could be true. The slides might not all make sense, but the team is fantastic, the downside is not that big, and an angel investor is willing to bet on that one big assumption.

Photo by dylan nolte on Unsplash

Slidemagic 1.0 templates merged into 2.0

I am working hard to get rid of the Shopify template store with its shopping carts that are great for buying t-shirts, but not convenient for downloading presentation templates.

As a first step, I have now merged most of the slides of the Shopify store into the SlideMagic 2.0 database. Beta testers who go to the web site, see the templates alongside the new ones generated by the app without noticing the difference.

For 80% of the slides, I could easily convert them to .magic (the boxy ones). These slides appear in the template store with both a .magic and .ppt download option (the .ppt conversion is generated by my software, rather than the manual adjustments for screen size etc.). For the other 20%, I have uploaded the .ppt file without a .magic option. If you are just after .ppt downloads (hopefully you will change your mind at some time), you can use the template store without noticing the difference.

Users that access the slide database from within the SlideMagic app will not see the PowerPoint-only options.

So, hopefully I can retire the Shopify site soon, and I will migrate the subscribers to the new site. And, this exercise gave me some insights into what shapes I should add to SlideMagic, and what shapes are actually not required at all.

This is all a bit of a ramble by a product manager who is trying to justify and integrate past product decisions :-) What it means in practice:

• You signed up for the SlideMagic template store in the past: you can access the same templates, but also new ones

• You signed up for the new SlideMagic 2.0 app: you will have access to a great set of templates that are super easy to customise and don’t even miss the PowerPoint-only ones. (And, all your work can be converted to PowerPoint at the press of a button)

The SlideMagic 2.0 template database is now already a lot bigger than the 1.0 one. To be continued.

Beta: related slides

For sites such Amazon or Netflix, ‘related products’ searches have become reliable and useful. On my Shopify platform it was very difficult to implement a link to “related slides”. Now that I have acquired the required coding skills, I looked around at other out-of-the-box search solutions (even with the help of some suggestions by blog readers), but I have not found something that could work for me in SlideMagic 2.0.

I think a good search algorithm is absolutely crucial for a slide template engine. Simplistic tag searches are not useful. “Here is another slide about "‘strategy’). You need to take into account how a slide looks, what message it should carry, what audience it is relevant for, etc. etc. And then in addition, bring in how popular a slide is.

I am breaking my head on finding a good way to index slides and make them accessible in search, both on the web interface, but more importantly inside the desktop app.

Today, I completed a very first version to get the basics working. Still far from perfect, but already better than the Shopify site. Next up is a refinement of the tags, the tagging structure, and the user interface. And of course: getting the performance right.

This has not been incorporated in the desktop app yet, I am using the web site as a playground for ideas at the moment.

Work in progress.

Any Office add-in experts out there?

I am trying to create an add-in that brings the SlideMagic 2.0 template store right into PowerPoint: search for slides, “click”, and the layout appears in your document. I thought that new Office APIs would make this really easy. I have a web server already up and running, the only thing is to move to the task pane at the right of the PowerPoint screen.

Googling around, the tricky bit seems to be the final step: inserting the slide template into the user’s presentation. The PowerPoint version of the Office API seems to be really limited. The furthest I can push it is to open a brand new presentation inside PowerPoint with the selected template(s) in it, the user then needs to copy them across as a final step.

Maybe one of my readers can point me to a solution? (Or maybe encourage Microsoft to expose this feature in their API, it will open up a whole raft of possibly very useful PowerPoint plug ins I think).

Obviously, I can dive into the world of hardcore manipulating of XML/PPTX files, but I am not sure whether that investment in time and effort is worth it.

Photo by Randy Laybourne on Unsplash

Boxifying...

The philosophy behind SlideMagic is to save making everyday presentations:

• A bank of useful templates (still small, but growing rapidly now)

• A clever search algorithm

• A proprietary UI to make obvious changes to slides at lightning speed (with the option to convert to PowerPoint)

• A simple framework to make all slides look uniform and consistent with corporate branding

But there is one other important component: simplifying (call it “boxifying”?) layouts.

Shapes and layouts that are great for drawing on paper or a whiteboard, are harder to get right and look good on a computer, think circles, arrows, curved lines. Professional designers know how to space text out evenly, add white space, line up the 7 levels perfectly on a circle at even distances. For the rest of us, this is a lot harder, or maybe even impossible. Not everyone has that eye for design, you know your slide looks bad, but you somehow cannot pin down why.

Another problem with these “sketch shapes” is that they are not ver efficient to hold text. Try adding long words in a circle shape and you run into problems. (The Japanese figured out a way to create square water melons so more of them would fit in a fridge).

And maybe you have that eye for design, then you still need to invest all that time to get your shapes and layouts right. Time that is worth it if you are designing your annual sales team kick off keynote address, time that is definitely not worth it when bashing out the quarterly numbers for a quick review meeting with the team.

As an example, see this alternative layout to the classical Venn diagram that I recently added to the SlideMagic 2.0 slide database. Less pretty, but it still looks decent and is easier to work with.