This is part 2 of a two-part series on corporate startup collaboration. The first part was a tongue and cheek guide on how NOT to work together with startups. I hope you have read it and found it interesting. If not:
Now to the more serious business of what we have seen working.
Step 1: Establish first contact – It’s easy
In the beginning of a collaboration, there is the first contact. For people unfamiliar with the matter this might seem like the most difficult step. But actually, it is not. It is almost easy. You as corporates are interested in innovation, we as startups are interested in gaining clients.
There are many accelerators1, VCs or venture-builders2 which intermediate through events and startup collaboration formats. It is its own industry and they can add expertise in scouting pre-selecting startups which you don’t usually have in-house. Larger companies have their own programs3 where startups can apply. In addition, we at GenLots actually book most of our meetings through 1) our own intermediaries which have a deep network in the industry and 2) cold outreach through email campaigns.4
What is important is what comes after you have booked your first meeting.
Step 2: Assembling the puzzle – your biggest job
Once contact is firmly established, the next step is to assemble a puzzle internally. It consists of:
- Assign Point of Contact (see also step 3 below)
- Find a Business Unit willing to dedicate time test the product
- Validate with end-users / domain experts
- Position the solution within the current software landscape
- Get your IT on board
- Approve (& negotiate) budget
- Ensure C-level support
- Establish a clear timeline
I think most of the points are pretty obvious, but one which I want to stress is the positioning. If a startup is doing something highly specialized and niche such as GenLots, it is important that stakeholders can visualize where in their huge puzzle of existing softwares we fit. We have to show that we are impactful, unique as well as complementary. This is easier said than done. Many SaaS solutions act on inventory reduction. Why is GenLots different? We are convinced that we do something nobody does and that we work well with other solutions. But how to put it into words, into the right language, such that an audience of 5 wildly different people and roles at a corporation can understand?
The second aspect which merits a little aside is timing. When we contact your company at the right time of a cycle, we can become part of a target architecture, integrate with an existing project you were planning anyways. If we are coming in at the wrong time, you are already doing 5 projects in parallel and don’t have the resources to analyze anything in addition. Actually, your company is probably so big that you are actually in both situations at the same time, just in different geographies or business units.
So our most successful projects either 1) attached to an existing project (i.e. “One MRP across all regions”) or created an entirely new one which was easy to explain in one sentence, aligned with internal objectives (i.e.”get to order planning excellence”).
It is difficult as a start-up to assemble the puzzle for each individual client, especially early in the process as we are still learning how the organization works. It is like translating a text into a language we don’t know very well – without AI. So the maybe surprising conclusion is that the biggest part of assembling the puzzle has not to be done by the start-up, but by you, the corporate!5 Only you have the right vocabulary, the right network and know who to talk to. And to do this right, you need great people!
Step 3: Put the right people on the task
It is important for a corporation to do its due diligence when considering SaaS solutions. Information Security is paramount. IT has to check if they cannot do this internally or with an already bought software package. Procurement wants to see at least one alternative (even if the startup considers itself unique) to make sure that you don’t buy something which they could get cheaper with more features elsewhere.
Therefore, chances are that you have many people who worry about the due diligence and the default to any new software is no.
So to assemble the puzzle above you need someone who can quickly assess the impact of a new solution, validate it with key stakeholders and then, if really convinced, drive this project against the inevitable headwind. Such people are well connected, have power (or know someone who does) and take ownership of the project. And such people are very rare but exist. If there are none at your company, chances are very low that you will adopt new SaaS solutions.6
Step 4: Do a test phase instead of a pilot
In part 1, I mentioned that a pilot can really lose you momentum. But validation is still important. So instead of a pilot, do a test phase. What is the difference? A test phase is a production implementation with X months of testing at the end and an easy way to abort if it doesn’t work out. A pilot is a general test of whether the startups’ SaaS software brings value and whether its tech works with some data input of your company. A little caveat: We are a software company, optimizing an existing internal process. This part can be fundamentally different for other types of startups.
The advantages are manyfold of doing it this way:
- Just doing it is the best way to test: In software development, it is easiest to validate a concept by doing it instead of talking about it. I am serious! At GenLots we had a case where a client migrated from an old ERP version to a new one. We as GenLots had to adapt. For this we had to participate in around 30 hours of meetings with an average of 5 people from the client (so 150 hours of effort for them, 30 hours for us) just to take in the end 8 hours for implementation and testing. Important to mention: around half of the effort of this one day we spent on problems not foreseen in the 150 hours of planning! Analogously, we found that you can talk a long time about implementing GenLots, but it actually only takes several days of effort to do it.
- You have tested the hardest part: The hardest part is not validating the business value7 nor the technology8, but whether the start-up works with your specificities and whether the day-to-day use is practical.
- You can still start small: Do it just at one site, one region etc. and then scale up. But do it end to end with all the hard parts in the middle such as integration.
- The effort is higher than a pure pilot, but at the end you can directly go to production without any more changes, whereas after a pilot you still have to worry about potential roadblocks for implementation, usability concerns etc. So you are almost back at square one.
At the end of the implementation, you can test the software several weeks in production and then either go for a license or stop the project.
Along the way: Plan for internal objections
There will inevitably be internal objections. For handling them we have seen several best practices which help:
- Accept the objection. Besides the objection often having merit, it is always better that they are voiced and you can address them. Otherwise they remain in the head of the stakeholder and cause trouble down the road. We have seen that if you can convince the stakeholders who objects most vehemently, you win the project.
- Validate the impact of the objections: Often with GenLots, objections concern specific materials to plan which have special planning procedures in place which cannot be optimized with our algorithm. But if 90% of the other materials work with our algorithm, maybe we can agree to look into those after we have saved millions of dollars?
- Not today but later: Similarly, many objections don’t have to be addressed from the beginning. If you can split a project into several stages, where each stage is “feature-complete” and brings value, you can shift objections into later phases. An example: GenLots offers an API-only model where we suggest your next orders and we offer a policy advisor where we, among other features, optimize safety stocks. Many companies already have tools in place for safety stock optimization. So instead of initiating a lengthy comparison with those tools, maybe we can just agree to leave out this feature for the moment and already concentrate on the core feature of GenLots, order optimization, which is expected to generate a lot of value in itself and has no competition.9
- Adapt: One of the main reasons you wanted to work with a startup like us in the first place, has probably been that we are known to be flexible and fast. This is true. Probably much more so than you can imagine when you come from a corporate world.
A good example of how to voice objections from our early phases was when we were asked about IT security10 by a large corporate we did a pilot with. The sequence was as follows:
- We went through the regular security assessment. We responded very honestly.11
- They identified deltas. In our case that we didn’t offer Single Sign On12 and that our security was not assessed by an external party
- As part of the implementation, we added Single Sign On paid a modest amount to a ethical hacking company to get red teamed13
The effort to add the additional requirements was actually not that big. The clients’ IT organization was happy and we came away feeling more enterprise-ready with new features for future clients.
Those are just a few of the ways on how to address objections.
Where that leaves us
After a rollercoaster of seven years at GenLots, we have learned a thing or two about the dance of corporate-startup collaboration. Our naive optimism at the outset clashed with corporate realities, but was essential in continuing to improve. There are two types of innovation programs at large corporations. The one from part 1 of this article which makes you feel innovative, but actually is a start-up zoo and initiates pilots which resemble more corporate limbo than a launchpad. And then there are the ones who actually adapt their processes, are ready to send in the right people, assemble the puzzle masterfully and have a good way of voicing and addressing the inevitable objections. The latter programs bring an undeniable competitive advantage to the company and can create proper innovation factories.
- Specialized in corporate-startup collaboration we participated in: Plug and Play, MassChallenge, SAP.io
- Just some examples: Pioneers (Austria), Rainmaking (Global), Co-Cubed (London)
- Example Unilever
- They work surprisingly well. Our best campaigns get over 50% open rates from C-level executives we don’t know at all.
- Or, if you are lucky, you have a commercial partner who knows the organization very well to accelerate positioning immensely. Nowadays, when we have issues positioning GenLots, we try to get the account manager of our partner SAP for this client on the phone. He will know much better than us, and maybe even better than our champion, how to position GenLots.
- Adoption speed also depends on the department. Marketing and sales or IT is a much faster adopter than i.e. supply chain….
- We at GenLots have a way of easily showing on the companies own numbers what the potential savings of implementing us will be – called proof of savings. But we stopped proposing that. Why? Because we did many of those proof of savings without any sales in the end. We didn’t understand that business value remains always open to criticism before you have delivered. You can evaluate the business value differently than with a pilot: Talk to our existing clients, do a back of the envelope calculation with experience values given by us. This is much less effort than a pilot.
- This is maybe bold. But if a startup is not a complete fraud, their tech has already been tested by themselves and ideally by a first couple of clients. And you probably have no one qualified to test the tech anyways. In the case of GenLots, a reinforcement learning algorithm for cost optimization.
- We really don’t
- As a very small, early-stage enterprise SaaS startup, IT security is at the forefront of all your thoughts. At the same time you don’t have the budget for costly ISO certifications. So when you hear the name “security assessment”, you hear the tune of darth vader from star wars in your head
- We also somehow benefited from our small size. Example question from IT: Are security roles and responsibilities for every role (Including Non-IT/Security) defined, documented and signed? → Our answer “not right now but they can be this afternoon”.
- This is a feature whereby i.e. the employee signs into his windows machine and is then automatically signed into all the services, without having to have separate passwords for software such as GenLots
- One of the best tools to assess security. A team of employed hackers try to get into our systems and give a report in the end with suggested security improvements. Also called pen testing