2013-05-14
New features on the startup page

During the last days I completely revamped the startup section of my blog. Here are the main changes:
- now rather than just a short text and a radar chart you’ll find detailled tests of startups/products (see my last test of intercom.io)
- you can now share you own short review for each product. All you need is a Twitter account to login and then to fill up your own review form which is composed of a simple text box and rating fields to create your own radar charts.
If you disagree with me you can now just share your point of view directly on these pages :-)

As usual if you have any feedback don’t hesitate to share it with me.
2013-04-26

Two cool demos from Algolia: search the Appstore and Crunchbase

I recently wrote about Algolia a promising web startup which provides Search as a Service. I was really impressed by the product and I really think that they are onto something.

The team recently put together two demos to show what you could do with Algolia and both are impressive.

The first is a crunchbase search which uses their javascript client. They have indexed crunchbase content (more than 270 000 entries of startups and people) so you can search it directly through Algolia technology.

The second demo enables you to search the Appstore with more than 630 000 apps indexed.

First thing I noticed was the speed of the search, it’s really fast. It takes only some milliseconds to get your results. It was one of the questions in my test: how efficient and fast is Algolia when the index has hundreds of thousands of entries. Well it’s pretty fast.

Second point the javascript client (both demos use the javascript client) makes it easy to integrate a search on your website and to provide a really good user experience: typo handling (especially cool on mobile), displaying results after each keystroke, highlighting the part of the result which corresponds to the query etc… All of these features easily done once you have indexed your content on Algolia servers.

Because of Google we are now used to expecting all these features from a search (typo, highlighting, results automatically after each keystroke etc…) however it’s far from every big website and mobile app that provide this user experience. Now with Algolia it just gets easier for developers to do so.
2013-04-19

You are probably not the only one to have this idea and probably not the only one to come up with this solution

Yesterday I discovered Bitium a SaaS tool that helps companies manage better their SaaS apps and user passwords.

What is funny is that the concept of this product reminded me exactly of what we were working on on my previous startup (which stopped). I immediately tried Bitium and the product itself has many features that we had in our prototype, features that are almost identical (like the onboarding or the SaaS marketplace), even from the UI point of view.

Of course the point is not that they have copied us (they didn’t know us at all and I discovered them yesterday), the point is that when you take a bunch of people and make them think hard about a problem (in this case SaaS management) you’ll probably end up with very similar solutions at the end, whether you do it from France, from Los Angeles or Japan.

It’s known that when you suddenly have an idea which you think is original, chances are high that somebody on the internet already thought about that. But it’s not only true about the idea itself, it’s also true about the implementation of your idea, the product itself.

This is why when you see a trend coming you can bet that many startups will propose very similar products and approaches. Very often the first to launch will claim that the following competitors have copied them but I’m convinced that in the majority of cases it’s not true. It’s just that there are trends and cycle in the technology world and when many entrepreneurs start to be interested in these trends similar solutions organically pop up.
2013-04-17
Nexmo: a (cool) API for sending and receiving sms

Nexmo is a very cool API that enables you to send and receive sms. Let’s say you need to notify your users when something happens on your service and you want to notify them directly by sms rather than email only, Nexmo lets you do that programmatically.

With Nexmo API you can:
- Send sms. You make a call to the API by providing the phone number to send the sms to, the text and Nexmo automatically sends the sms
- Receive sms. You create a virtual phone number on Nexmo and when your users send sms to this number you get notify and can do whatever you need (perform some actions, send them back a confirmation sms etc…) (note that receiving sms is not available in France yet).
As you can imagine the advantage of a sms API lies in its flexibility:
- through the rest API you can easily manage all your sms notifications: send a confirmation message to your users, notify them in real time when something happens on their account etc…
- you pay what you consume, there is no prohibitive upfront cost. The price of a sms depends on the country where you send it but for example in France the cost is 0.045 euro per sms.
Like emailing, one very important factor for sms is deliverability. When you send a sms you want it to be received by your user. Nexmo is dealing with deliverability not only behind the scene with its technology but also by providing users with guidelines for each country.

From the API point of view the documentation is really clear with examples and libraries for several languages and platform (ex: Python, PHP, Ruby etc…) so starting sending sms can be really fast.

In conclusion sending sms through Nexmo is definitively something I would consider doing. I really enjoyed testing their API: it’s simple to understand and yet offers great possibilities. I just used it for a test (I received my test sms almost instantly) so I cannot speak about deliverability for hundreds or thousands of sms sendings however they show some pretty big customers on their homepage.

You can try the API for free since Nexmo gives 2 euros worth of sms to spend when you register.
2013-04-10
Bunkr : a promising presentation tool

Pitch

Bunkr is an online presentation tool which not only enables you to create your presentations (a la powerpoint) in your web browser but also to collect and store content (image, video, quote, text) from the internet which you can use later in your slides.

Customers

Bunkr aims at all people (mostly professional) who need to create presentation and slides.

The business model is subscription based. Like many SaaS you pay every month access to the tool, the pricing depending on your usage (how many users can collaborate). The tool is free during the beta.

Need

The startup is aiming at a market which is already well established: presentation tools. The need is simply: “help people create and manage their presentation online”.

Product

If Bunkr is going on an already established market then they have to approach it differently, which is the case: instead of going for the “best presentation builder” the startup wants to make easier for people to collect content on the web (images, videos, text, quotes etc…), content that you will use later in your presentation.

So one of the first thing you do after you register to Bunkr is to install their bookmarklet with which you will collect content.

The concept is pretty simple when you are browsing the web it often happens that you find nice infographics, beautiful pictures or interesting text, quotation and figures that you can use to illustrate a slide. Instead of saving them somewhere on your computer you just click on bunkr bookmarklet, indicate which content on the web page you want to save and it automatically creates a slide which is saved into your Bunkr account. The day you need this slide you just look for it in your Bunkr space and add it to your presentation.

Of course the other important part of Bunkr is the presentation builder. Now that you have content you can create directly from your browser classical “slides” presentation, collaborate on them with your team and finally export them or share them on the web.

From a product point of view it’s still a beta version so some features are missing and it needs some polishing in order to offer a first class experience:
- more options for the presentation builder. The tool is pretty limited at the moment and some basics features are missig (text formating, image editing, templating…)
- you cannot easily share content between slides. Imagine that you have collected a nice picture and some figures from several websites and that you want to mix this contents on one slide, then it’s not easy to do on Bunkr.
In conclusion for a beta the tool is already great but it still lacks important features to compete with existing tools.

Competition

As we said Bunkr has choosen to enter an already established market so the advantages are that users are already educated and they know that money can be made but the counterpart is that many players already exists:
- Powerpoint and keynote
- Google presentation
- Prezi, SlideRocket etc…
Clearly it’s going to be one of the hardest point of their mission: trying to be better than the competition and making users move from tools they are already used to, to a new tool like Bunkr.
2013-04-08
Too many SaaS ? Problems and opportunities

If globally the SaaS model benefits companies which use them, the explosion of the number of SaaS also creates some problems which impact directly the way businesses consume softwares. Problems mean also opportunities for third party players to solve them. We’ll look at 5 pain in the SaaS resulting this situation and how some startups are addressing them.
- Discovery: find the good SaaS which fits your needs
- Integration: integrate SaaS with your existing processes and tools (softwares, ERP)
- User management: how to manage users within your company (creating new accounts, managing existing ones etc…)
- Billing management: how to centralize SaaS billing
Licenses and SLA management: how to manage different types of licenses and SLAs

Discovery

With probably more than tens of thousands of SaaS on the market it’s getting more and more difficult for businesses to find the good one. If you look only at CRM softwares, for example, you’ll find plenty of SaaS which offer different features with different pricing and different SLAs. In these conditions it’s really hard to choose the best one which fits your needs and constraints.

More and more startups (and bigger players) want to address this “discovery” problem and their solutions come with 3 main flavors:
- marketplaces (ex: Google apps, Salesforce…)
- news website (blog, IT websites…)
- recommendation / comparison engines (Bestvendor, getapp…)
Discovery is a hard problem to solve which exists in basically every industry: internet has Google to help you find out what you want, online commerce with price comparison websites and marketplaces (Amazon, Etsy, eBay)…, mobile apps with appstores and app recommendation engines etc… It’s usually a hard business because it requires high volumes: you need a lot of items to attract a lot of users in order to make money (especially when the business model relies on affiliation). A lot of players are tackling this problem and it’s hard to know who is going to win… But it’s a high volume low margin business.

Integration

Integration is also another big concern for a lot of companies which want to consume SaaS. It’s great to use a software with a neat user interface and innovative features but very often it’s also critical that it offers bridges with your existing softwares and can be easily integrated with your IT environment and processes.

More and more SaaS makers now provide bridges to other applications through their APIs but it requires work to properly integrate them so there is still a lot of opportunities for third party players to solve this integration problem.

Startups like RunMyProcess, IFTTT, Zapier bet on “process” and “workflows”. Basically they help people create connexion between SaaS (and sometime on premise softwares) through drag and drop interfaces. For example you can connect Mailchimp with your Salesforce CRM through IFTTT.

Integration is a real hard problem, hard to solve from a technical point of view: there is no international “services interconnexion” standard widely used so a lot of customization is needed. Scaling is also another concern: if an interconnexion platform offers too few connectors it won’t be used by businesses so they need to have a technology flexible enough to integrate thousands of services and to keep up to date with possible changes. With the API trend this problem will probably get easier to solve in the years to come.

User management

More softwares as a service also means more user management to do:
- onboarding: how to deal with creating new accounts, setting privileges and teaching people how to use them.
- offboarding: how to deal with employees leaving the company and making sure they don’t have access to your data afterward.
- password management/security: how to deal with password security
Identity management has always been a problem in enterprise softwares with solution like “single sign on” that have been dreamed of for decades but are barely working. It gets even more complicated with so many SaaS, software marketplaces, devices (mobile, tablet, PC etc…).

A bunch of startups are tackling identity management and password management like Okta. If you look at their team and the funds they have raised so far (more than 50 millions of $) you can see that it’s not an esyl problem to go after.

Billing management

When you use plenty of SaaS within your company and when the time comes to pay the bills it can get quite complicated. We often hear that one of the SaaS model advantage is billing flexibility (subscription, pay what you consume…), however it can get quite complicated when you have to deal with many SaaS at the same time.

There are two main problems in SaaS billing:
- billing overview: you need to consolidate all the billing data coming from all your SaaS easily, centralize them in one convenient place.
- optimization: since SaaS billing can depends on your usage (how much you used, how many users etc…) it can be easy to pay more than what you really need, this where cost optimization tools can help.
A lot of startups are also trying to solve this problem. Like Cloudability which deals with both of the points above or Cloudyn which specifically look at your AWS usage to optimize its cost.

License and SLA management

Last problem I wanted to talk about in this article is SLA management. Having Spoken with a lots of companies that want or are using SaaS, one of the biggest fear they have is how to deal with SLAs.

It’s an important topic but difficult to deal with: how to distinguish what is important or not, how to know which SLAs you are really dealing with (ex you use a CRM SaaS which uses AWS, are you also depending on AWS SLAs?), can you negotiate your SLAs? etc…

I don’t know any startup that is trying to solve this problem (maybe it exists) and I don’t even if it’s possible but I think that it’s a real problem and there is money to be made here.
2013-03-21
Algolia: a promising Search as a Service API

I covered Algolia some weeks ago but didn’t have the opportunity to try their product. This is now done, I could test their Search as a Service API.
- Search is important (not for all apps but for a lot of apps). Users need to search and find easily content/data through your mobile app/website search
- Yet it’s complicated and hard from a technical point of view, good search is not an easy topic.
- Search is important but at the same time it’s not an area where you should put a lot of resources (except if your search is really the core of your product). I would compare it to emailing and payment: you can do it yourself but you should probably use a third party app or service.
Algolia provides such a service with a Search as a Service API and a SDK. I tested this week their Search API.

How does it work

The concept is pretty simple:
1. You prepare the data you want your users to be able to search on. Ex: for my startup list I wanted the users to be able to search on the name, description and tags fields of each startup object.
2. You send these data to Algolia via its API, the service then automatically creates a search index on their server
3. When a user makes a search on your website you just query your Algolia index via an API call which gives you back the results that you can serve to your visitor
4. that’s it!
Of course you have plenty of options to play with if you want to tailor the search experience to your needs:
- you can decide how results are paginated
- you can decide which attributes should be retrieved
- you can set typo tolerance settings
- you receive some ranking info for each result (convenient to sort your search results differently than the order provided by Algolia)
- …
You also have everything necessary to manage your search indexes:
- you can create multiple indexes
- you can update them
- you can delete them
- you can retrieve, update, delete specific object from each index
- you can edit individual settings for each index
- …
The advantages (keep in mind that I’m not a professional web developer) in my point of view:
- you don’t have to set up and manage a search engine in your backend. I did it once with Django Haystack/Solr and I really didn’t enjoy it. Hard to customize and I had some compatibility problems with other django apps. I did also use Google Site Search but you have less control on your indexes.
- it’s easy and fast to implement (it will be even faster when they release some clients for popular languages).
- Algolia handles typos, multi attributes search, result ranking etc… directly out of the box
- actually a weakness of my test is that I use it on my website but Algolia benefits are even greater on mobile apps, especially thanks to the typo handling (ex if you search “pqris” instead of “paris”) since typos are much more frequent on mobile. There are even more features with their SDK: Autocomplete, document search etc…
As I’ve said above I’m neither a pro developer nor a search specialist so I cannot give opinion on some points that would be worth discussing:
- performance and speed. From my tests it’s really fast, however my index is ridiculously small (12 items with 4-5 attributes). How does it handles indexes with millions, tens of millions of objects? (note that they also offer an “offline” SDK)
- results relevance. Again the relevance of the results is really good on my index but for some particular cases (very long text or plenty of attributes for examples) I’ve no idea if the results are still good.
- how the startup will handle scalability
- probably many other points that I’m not aware of :-)
In conclusion I’m really enthusiastic after by this test and I really think that Search as a Service is interesting in many cases. I will definitively use it.

Keep in mind that the API is still in beta mode so it lacks documentation and tutorials as well as some clients (for python, php, ruby etc…). The pricing for the API usage is not out yet. If you want to try it contact directly the team.
2013-03-20
B2B software marketplaces

We’ve reached a point in the software/SaaS industry where marketplaces are becoming more and more important. Exactly like in e-commerce where the ecosystem is composed of million of independent websites which orbit huge marketplaces: Amazon and ebay (and also some smaller vertical marketplaces: ex Etsy for homemade stuff)

The situation is the same in the software industry. With the SaaS model it’s never been easier to launch a software, there’s probably hundred of thousands SaaS out there. The problem are the same as for e-commerce: how to find the good SaaS and consume it easily?

Loose or controlled marketplaces?

Of course marketplaces come with different flavors and we cannot list all their characteristics, so I’ll focus on one of the most important: how the marketplace operators operate their marketplace.

There’s two main philosophies: be loose on control deeply your marketplace.

The “loose” philosophy lets the users and sellers “free” on the marketplace. It’s easy to put an item on sale and you have many choices for the payment method (credit card, escrow, cash…).. The operator doesn’t enforce many rules on his marketplace. eBay started as a very “loose” marketplace, allowing all kind of items to be sold and users to pay cash. They since backed up to a more controlled experience. Leboncoin (a french marketplace) is another example of a “loose” operator, it really doesn’t control any things on its platform and are very successful on the french market.

The other philosophy consists of controlling much more what is going on on the marketplace: strict rules for which items can be sold or not, controlling the sellers beforehand, enforcing a unique payment system etc… Amazon is a typical example for this kind of marketplace, everything is controlled in order to give the buyers the best experience possible.

Well it’s the same for software marketplaces you can find marketplaces that act just as “apps repositories” where it’s easy for sellers to be listed and where the buyers have to figure out by themselves how to buy the SaaS and you have software marketplaces that are like Amazon and control every part of the marketplace.

If we look at mobile apps marketplaces we can see that Apple has chosen to control every step: sellers have to pay to be there, each app needs to be reviewed in order to appear, some categories are forbidden (ex: porn / adults apps) and the user has no choice in the payment method, it’s via Apple and nothing else.

Android on the other hand offers an intermediate experience, it’s not completely loose but it’s still clearly not as “controlled” as the apple appstore.

Each marketplace model offers their own advantages/drawbacks but I won’t come back on it in this article.

B2B softwares marketplaces

Now let’s come back to B2B SaaS. As we’ve said first, there’s more and more SaaS appearing every week and it’s getting more and more complicated for enterprises to find and consume SaaS (and for SaaS developers to distribute them).

I think that SaaS marketplaces will get more and more important in the future. Of course there’s nothing new here since Google already offers a B2B marketplaces with Google Apps or SalesForce which is one of the dominant players with force.com, but the industry is still in its infancy and will only grow.

What is interesting with B2B SaaS marketplaces is that they are conceptually the same as B2C marketplaces but they have to offer different services to the users (the enterprises). When enterprises want to consume SaaS they face a lot of problems:
- storage: where are the data stored?
- legal: under which laws are my data covered?
- technical: Is there single sign on? Connectors?
- SLA: how can I deal with different SLAs? Does the marketplace provide standards?
- etc.
Marketplaces could be a great model to ease these pains by offering a controlled experience on their platform. Example: SLA (Service Level Agreement) are a pain to deal with. Each SaaS has its own, so a company which uses 10 different SaaS potentially has to deal with 10 differents SLAs which is a big problem. If a SaaS marketplace enforces a unique SLA on its platform it could be a great help for consumers.

All these problems are not fully addressed by the existing marketplaces and even the way marketplaces will be used within companies is unclear at the moment:
- Are small companies more likely to use public marketplaces like Google and Salesforce?
- Will big companies prefer more closed or private marketplaces?
- How can SaaS marketplaces provide the best experience for companies?
- Is the marketplace ecosystem going to evolve like ecommerce: a dominant players and many vertical smaller players?
Still a lot of questions to be answered.
2013-03-18
capitainetrain: buy you train tickets easily

Pitch

capitainetrain sells train tickets online.

Customers

Everyone who uses trains. The business model is simple: they take a cut for every sell they make.

Need

The startup answers two problem:

1- Generally ecommerce train ticket websites are complicated, full of ads and try to cross-sell you plenty of other things (hotel nights, plane tickets etc…). capitainetrain wants to offer a user experience focused on simplicity and speed: buy easily your train tickets.

2-Generally ecommerce train ticket websites are owned by the train operators themselves (SNCF in France, Dutch ban in germany etc…) which makes it hard to buy tickets for different countries. Ex: I want to do Marseille → Paris then Paris —> Cologne and finally Cologne → Berlin. You can buy all the tickets through SNCF website but prices are not optimized. It’s usually better to buy your tickets directly from each operator website, which is really inconvenient. capitainetrain wants to solve this problem by offering a central place where you can buy all your tickets with the best prices.

Product

In the french startup ecosystem capitainetrain is seen as a very promising startup with a great product oriented approach that focus on the best user experience possible.

So I was quite excited to try the product to buy my train tickets from Germany to France but I was a little bit disappointed at the end.

First I have to say that I really like the promise of a website which focus on what I need to do (buy a train ticket) rather than on trying to cross-sell me dozen of things and making the user experience crappy on purpose in order to display more ads.

From this pov capitainetrain doesn’t lie. The product is minimalist (it’s not negative, I like it actually), the interface is really focused on selling you train tickets. The first thing you see is a form with your departure and come back dates and of course the location of departure and arrival.

By filling up this form the website gives you the different trains available and the prices for each. Once you have chosen your tickets you can pay directly from the app. The process is pretty slick.

The design of the results is really “compact” and gives you some data to help you make your choice: of course prices and time but also comparison between first class and second class, type of ticket (flexible or not), duration…

However this approach of “the simplest product” possible also limits the user experience. I was frustrated by two points:
1. it’s hard to find at a glance what is the best price for the day and it’s impossible to see the best price for a range of dates. When you don’t travel for business purposes you are generally more flexible on the dates and would like to see what is the best price within 2 or 3 days. It’s especially true for train tickets where you can find great price differences between 2 consecutive days. You cannot do that on capitainetrain.
2. when you change the date of arrival or departure you have to complete the process all over again. Let’s say I want to leave the 04/02 and come back the 16/02, I found a good ticket for the departure but not for the come back, if I change the date of the come back I have to do all the process for the departure again.
The last point which frustrated me was not related to the UX but to the fact that I couldn’t buy the mobile version of my Thalys train tickets. With the Thalys you can choose either to receive your ticket on your mobile phone (and by email) or to have the regular version that you have to print from the train station. If you don’t want to queue at the train station only to get tickets already paid the mobile version is really cool, however I couldn’t get it from capitainetrain which is a big problem for me since I could do it directly on the Thalys website.In the end effect I had to buy my tickets on the Thalys website.

I don’t know if this limitation is due to the Thalys system or if it’s just a lack from the startup website.

Competition

Competition is tough on this market. voyage-sncf is one of the biggest ecommerce website in France and has quite a lot of money to spend on marketing. However this is also true that their website is far from offering the best user experience possible and in this perspective capitainetrain has its chance.

Conclusion

I really like the concept of capitaintrain and its philosophy but I was a little bit disappointed by the lack of key features (for my usage). Maybe I was expecting too much from a startup from which I already heard so much praise. I hope they will offer more features soon without sacrificing their philosophy and then I will become a real customer of capitainetrain.
2013-03-14
Cinemur: an innovative web app for movie lovers

Pitch

Cinemur is a web app that let people discover the latest movie news, trailers, reviews through a very innovative interface which heavily relies on social features (Facebook open graph). Cinemur also adopted the same approach for movies that are broadcasted on TV.

Customers

The french startup main targets are you and me, basically people who are interested in new movies and want to know which ones they should go to. The service is free for these users (around 120 000 registered users).

Unlike the vast majority of “ traditional” movies websites, Cinemur doesn’t monetize itself through ads (which they don’t want on their platform) but through a B2B strategy with movie theater owners and movie producers for whom they offer viral promotion tools (Facebook). The app also monetizes its API.

Need

For you and me: discover latest movie news, review them and share your opinion with your friends and the Cinemur community. Same for movies that are broadcasted on TV.

For movie theater owners and movie producers: promote efficiently their movies through social features (Facebook and Cinemur community).

Product

The product is definitely the main strength of Cinemur. The app is highly visual and the result is just gorgeous.

The first thing you notice by going on Cinemur is the wall composed of movie posters. It’s through these posters that you access each movie info: the trailers, comments and reviews from the community and general info on the cast, some ratings etc…

What is interesting about this wall is also that it’s tailored to your taste. Thanks to its algorithm which gather the data you let on the website (what you like, what you don’t like, your friends taste etc…) Cinemur displays different movies and promotes some that other users won’t necessarily see.

The app offers many features
- discovery features: recommendation from your taste but also from the community, latest movies, movies to come…
- social features: reviews, friend list, facebook sharing, movie list…
- organization features: movie theater timetables, invitation between friends…
What is really great, from my point of view, is that you don’t think so much about how to interact with the product even if all the user interactions have been rethought compared to more “traditional” movie reviews websites (such as Allocine, IMDB…).

Competition

By starting a project which tackles movies recommendation and database the team didn’t choose the easiest market. This market is dominated by old “institutions” such as IMDB and Allociné which trust the top results of search engines. It’s basically impossible to enter this market through a search engines strategy, this is why Cinemur is betting on social networks.

Cinémur v2 - Présentation vidéo from Cinémur on Vimeo.

The startup not only offers a radically different product but also a radically different traffic strategy by favoring sharing between users and viral loops.

What I like

To sum up:
- really cool product both visually appealing and blazing fast. Even if the interface is surprising at first you quickly find your way.
- an innovative approach from the business pov by not having ads as main BM but by providing promotion tools to movie producers. This strategy could really work if Cinemur becomes a real platform for users to discover, review and share the movies they like.
- One of the few interesting and promising “social” startup in France