Guest Post by Pierre Fournier

It’s always special when we can feature a client on our point. we’re pleased to have a great guest post from Pierre Fournier who is chief product officer at Mano Mano. As a leading business for DIY consumers across the UK and Europe, ManoMano’s point needs to perform, scale, and help buyers and merchandisers find one another.
moment’s composition from Pierre first appeared on Medium. Please partake in Pierre’s wisdom and perceptivity with your network, especially as they consider erections vs buying new technology.

With the rise of SaaS editors in the tech world, you now have a tool for any need you can imagine( search, business, converse, content operation, payment). It can be an inconceivable accelerator if you pick up the right tools but also turns into agony when effects go wrong( migration, integrations, data power). So the “ make or buy ” decision came one of the most critical that a Product director can take in his job.

Then’s a summary of the criteria we assess at ManoMano( MM)

• Data power belongs to MM and isn’t mutualized to ameliorate the tool
• Commoditization of the point doesn’t make it strategic for MM
• Customization allows us to tweak the tool to MM’s specific redundant requirements
• Backward comity keeps us free to review our choice
  
• The ability of the editor allows us to have a long-term collaboration
• Costs shouldn’t be exploding if the operation grows

As a perk, we also partake in some tips to successfully run an RFP( Request For Offer) when having to elect a tool for a steal strategy.

Preliminary thoughts about the “make or buy” strategy

Make or Buy opinions can be really tough opinions, and I suppose that it’ll be more and more the case in the future. On the morning of my career, you hardly had SaaS tools( not to indeed mention interoperability issues) and you had to develop on your own nearly everything. There are many studies about make or buy opinions in the light of 10 times spent in technology attending buzz, trends, and sharing to RFPs( Request For Offer).

• Is there any new “make and buy” option? The rise of D2D( Developers To Developers) introductory technology layers will continue to grow and offer indispensable to pure “ off the shelf ” results, opening a “ make AND buy ” new path. For case Twilio offers the foundation bricks to make communication tools like a converse.
  Developing an in-house result can still make sense only a many range of third-party tool features are generally used( let’s say 20, it can indeed be less). That’s why developing an in-house result can remain a possibility, indeed moment. Let me give you an illustration I set- up a food tech company to allow online ordering. It took us 18 months to make a decent result. One of our associates erected a perfect result for an eatery proprietor in lower than 2 weeks other criteria championing for making a decision is the cost when the volume becomes big, off-the-shelf tools might not be as intriguing.
• Your plutocrat will finance substantially marketing and deals in a 3rd party tool, an important share of the costs are generally used to finance the editor’s marketing and deals You have to be apprehensive of that. Zoho offers a great illustration of a company investing a lot in products rather than in deals and marketing( compared to American mammoths like Adobe or Salesforce). According to Zoho, over to 50 of the earnings can be allocated to deals and marketing Investing this plutocrat into the product can make a big difference in the long run!
• Stylish- of- types strategy can come more applicable by having several technical tools rather than a fit each is another decision on which people generally spend a great deal of time. Keep in mind that big editors( generally the bones offering each-by-one results) generally develop their products through external accessions. Integrating the recently acquired products within the being product suite is a veritably hard task taking a long time Which occasionally noway happens! So each- by-one result might eventually turn out to be a medium best-of-types volition. either, APIs are now a standard easing the integrations of several products( likewise if you use data capitals like Member).
• No tool is a perfect fit to the tool rather than getting it to fit your own requirements, always complaining that this isn’t exactly the way you would have done it. else develop your in-house result! Keep in mind that editors, beyond their software moxie, also developed a process moxie through the dozens of guests they have. By paying a third-party tool, you should also profit from this process moxie.

Criterion 1: data ownership belongs to ManoMano and is not mutualized to improve the tool

This is presumably one of the most strategic criteria. MM data going through the tool shouldn’t be used to ameliorate the editor’s tool and potentially our challengers ’ performance. The tool’s value should come from the algorithm itself or from the features the tool offers.

For case ManoMano uses Algolia to power its hunt. We chose Algolia for two main reasons. Their applicability algorithm will always be better than what we’re suitable to do as of now and doesn’t take advantage of our data to ameliorate. Also, their interface makes it way easier for inventors or Product directors to make changes to applicability rules compared to Elastic Search(the former tech mound we used for our hunt).

still, we turned down another SaaS result offering query clustering because their algorithm would have taken advantage of our query data to ameliorate their own clustering and that would have potentially profit to our challengers.

Criterion 2: Commoditization of the feature does not make it strategic for ManoMano

When MM’s requirements are the same as everyone differently in the assiduity, there’s no value in redesigning the result since it won’t bring redundant value. We prefer allocating our tech coffers to developing secerning bricks of software. Again for Algolia, applicability, or rather document hunt, follows the same rules in every assiduity on a semantic base( you will see in the coming paragraph that behavioral applicability is different in our DIY assiduity).

A counter-example for us can be the CMS( Content Management System) that we chose to develop internally( indeed if we resorted to the low position being bricks like WYSIWYG editors, places, and administration operation). It’s substantially used to give Tip wastes to help our druggies choose the right product. As we wanted to connect it to our roster( product suggestions), to our taxonomy( to exercise the attributes we created in our pollutants, for case to explain them right from the tip wastes, and allow our druggies to constrict the range of products as they go through the tip distance).

You might have to review the request regularly because changes be at high speed. For case we were in advance to offer converse advice through a community in 2016. No satisfying results were available at that time so we took the decision to move on internally on a devoted platform. 3 times latterly, the request has progressed and out-the-shelf tools are now available. How hard it’s to go backward, we had to make this decision because shifting to a 3rd party tool would dramatically speed up the release of crucial features. And it would allow the platoon to concentrate on further strategic redundant features on top of these introductory features. This leads us to the coming criterium customization.

Criterion 3: Customization allows us to tweak the tool to fit ManoMano’s specific extra needs

A crucial point we’re looking at is our capability to tweak the tool. Generally, 80 of our requirements are participated with other challengers from our assiduity and can be addressed by an out-the-shelf tool. But there frequently remain 20 of the requirements that are specific to us because of strategic considerations. So we must still be suitable to develop these specific features on top of the 3rd party tool. It frequently relies on the use of our data to epitomize stoner experience( hunt, advice). The capability to manage programmatically the tool( via an API) is therefore critical. Algolia for case allows us to stamp the native ranking algorithm through an API enabling us to serve substantiated results grounded on behavioral data. Iadvize offers the occasion to customize native contraptions with ManoMano’s data(eg. product recommendations) or to add our own billing system on top of the converse structure.

Criterion 4: Backward Comity Keeps us free to review our choice after many times

Backward comity is an essential criterion because effects can change veritably presto( see our converse illustration over). New technologies can produce step changes( see databases operation for the case when Nosql technologies appeared) that new editors can profit from to produce big differences in little time. That’s why you need to be suitable to open a result at a reasonable cost( it always requires sweats). You need of course to be suitable to get back all your data( because this is a critical asset that takes time to rebuild, see the data power paragraph) and you need to limit as important as possible developments devoted only to this specific tool. The more you spend time and energy jotting law devoted to the tool to customize it, the least you’ll be suitable to throw away it. still, if you make custom in-house APIs to tweak the tool, you should be suitable to fluently throw away them modulo many updates on the transfer formats. Again in our example with Algolia, if we’d like a tore-internalize hunt, all the laws we’ve erected about our query-reranking could bere-used.

Criterion 5: The viability of the editor allows a long-term collaboration

When your company reaches a critical size( ManoMano is the online commanding platform for Do It Yourself in Europe), you have to be sure that your mate will still be there within 3 times. Because you’ll invest time and energy in integrating its result. In resemblance, VCs fund several tech companies every month that will come and see you pretending they’re the coming big thing, bringing you an insane competitive advantage. Some might be, but maturity will be dead within 2 times.

Does it help you from working with youthful startups developing cutting-edge technologies but not having set up scale yet? I would say you can, but only if

No other volition exists in the request(eg. we chose to work with a youthful incipiency on remitted payment because no other editor was)
Backing from VCs is strong enough to insure at least 2 times of development( I recommend the company to be at least in a B- series founding position)

It requires a veritably mature geste from both stakeholders. The customer mustn’t utilize the whole roadmap else the editor won’t be suitable to develop. The editor must be veritably transparent, suitable to refuse some features and advise if its profitable viability is in trouble.

Criterion 6: Costs Shouldn’t be exploding if the operation grows

Again, this item makes sense for bigger companies. As an incipiency you’ll have low volumes so bringing isn’t as much an issue. At ManoMano, our yearly business reaches nearly 50M of unique callers. Knowing that the utmost of the results is billed on a per volume base, it can come extremely precious at some point. This is generally one of the reasons a company will re-internalize a third-party tool. This doesn’t mean the result has to be cheap. Knowing that a Product platoon( PM, devs, contrivers see this composition for our rate) roughly costs around 500K€ per time, if the tool lets you spare the fellow of one or two features brigades, it means that you can pay up to 1M€ per time. To bring profitable value, it should be 30 to 50 lower. Also takes into account that doing it on your own requires conservation, that contrary to a third party tool you won’t presumably profit from the rearmost technologies updates, etc Accommodations with the editor are also keys, especially the thresholds when the tool is grounded on a per volume base. The further volumes, the less precious it should be hardly.

We had several conversations as well as respect to our PIM results( Product Information operation). PIM is relatively common to numerous diligence, features are always the same, and off-the-shelf results live. But eventually, we decided to make it on our own, because we thought it was a strategic asset( quality data product information) but also because the cost of the results would have been relatively high compared to the commission we bill to our merchandisers.

Bonus: How to run a successful RFP to select a tool?

When choosing a 3rd party tool, companies generally go through an RFP( Request For offer) process that explains their requirements, their current mound, and the features they need I would suggest keeping it as light as possible.

Then’s what I would recommend

1/ Experts should help final druggies, not the negative. Then’s how effects generally be a design director from IT, Product or indeed Procurement is entitled with the responsibility to run the RFP. He’ll conduct many interviews with some final druggies. And also he’ll write the decision grids, weigh each criterion, and interviews the editors I suppose it’s wrong. It should be the contrary final druggies should be at the heart of the RFP, supported by experts( engineers, product directors, design directors, and inventors). You should name a design director a person from the final druggies group.
2/ Involve as important people as possible in the RFP from every impacted machine. I know this sounds untoward and intuitive( we generally try to have the least persons involved to remain nimble) but by using for case Liberating Structures formats, you should be suitable. It will always save a lot of time by adding the liability that the tool fits with functional druggies’ needs Easing the steal-in of the unborn druggies( rather than a top-down approach like “ Then’s the new tool we chose, it’ll be great for you our lovely 150 client service agents! Let us explain to you why! ”)

3/ Try as important as possible to test the result through a POC. Deals rep from the editor frequently lies ( unconsciously) because they’re veritably far from the tech brigades, so you have to test the tool in real conditions. Some will argue it takes time. It does, but lower than enforcing the wrong tool. The way to do to minimize time spent is to reduce vastly the compass. By testing the result, you’ll snappily assess the quality of the specialized attestation, of the API, of the tech workers, the usability of the tool

4/Elect the 3 crucial criteria that would constitute a deal swell. Don’t get meddled in a hundred-criteria demand grid. You can fluently get trapped in an endless decision grid with weighted scoring. Indeed if it can make you feel reassured, it isn’t useful, because you can also miss the crucial point that would turn this choice into a big mistake.

Final Thoughts

Within every company, tech coffers are limited. On the negative, business ideas are measureless. “ Buy ” strategies can alleviate tech dependencies, especially on deals and marketing, and help businesses grow briskly. At least at first. However, it can come to a mess after many times and all the dexterity you first won can get lost into endless migrations and revamping, If not done wisely. So don’t accelerate in those opinions.

Besides all of that “Make or Buy” strategy, it will also depend on your company’s thoughts and ambition. Are you a tech company like Amazon? Or do you calculate substantially on marketing? ManoMano could have used an out-the-shelf tool like Mirakl to make its business. It would have been briskly. But when volumes get so high, the cost of the tool can come to a real issue. And we knew that if we wanted to make a commodity unique, we had to do it on our own.