Cooked. Luxms BI

There is no clear product development strategy and detailed roadmap. Essentially, the platform is a set of tools for creating analytical applications, which requires a large number of man-hours to implement the functionality. The lack of strategy is expressed in a sharp change in positioning; if in 2022 advertising went towards the development of self-service, then already in 2023, without having taken a leading position in this sector, the company begins to talk about the inferiority of the self-service approach. At the end of 2022, there was an announcement to repeat the Qlik associative data model; there was no continuation of this topic in 2023.

An attempt to launch a data catalog and a break from the principles of self-service will apparently lead to unsuccessful development of the interaction chain and another revision of the concept itself. As we now understand, the catalog itself has already become part of the ArenaData line, where the role of LuxMs will clearly fade away. Otherwise, the product itself will simply die without access to partners’ resources.

Focus on big business. Attempts to sell the platform out of the box end in failure and turn into protracted projects. The implementation itself is not possible by the customer, unless the client itself has a strong IT structure within the group. This is a big risk during implementation for both the client and the vendor. Apparently, several branches of platform support are already emerging, which may negatively affect the quality of the service itself and the timing of bug fixes.

Large projects are going extremely hard, as the vendor himself admits. Clients who purchased software in 2022 only achieved results by the end of 2023 through painful project delivery processes. This is largely evidenced by the summing up of the results by the vendor’s team, where there is not a single Customer online. Contracts for the purchase of software in 2022 are presented as new clients in 2023. It is not specified that we are talking about implementation, which the vendor is trying to take full responsibility for. This indicates that there is a difficult process of implementing projects.

If you look at the affiliate network, only heavy players are present there. The product is not large-scale, definitely not for medium-sized businesses. It is possible to obtain a positive effect (Value – TCO) only over a long horizon on a large number of users.

Based on the results of communication with a number of clients, we can note the low quality of technical support. The presence of bugs in the system is processed by the service with a delay. There are complaints about the timing of projects and the presence of postponements. Most of the comments are directed towards the vendor himself; partner projects are proceeding as normal. We can conclude that the vendor’s attempt to run an integrator business failed. By the way, this is a problem for many domestic players.

From the point of view of platform security, it is clear that the vendor devotes a lot of time to this, since he works in the upper segment of large businesses. But having a large number of product lines for each such large customer is a risk for the platform itself. Apparently there will be an emphasis on some kind of common connection of all components so that everything new works within this circuit. Perhaps they will end up in a more difficult situation with the portal, after the implementation of the necessary and sufficient minimum BI system.

Only in February we released online technical documentation for the platform! Most likely, this is a selection based on information from projects that went through the delivery process in 2023. The new release also adds mathematics, which is necessary for a mature platform. Most likely, for the next six months there will be a struggle to stabilize the new functionality and fix bugs. It would be interesting to compare the maturity of the internal formula language with the same DAX.

The ETL tool turned out to be very confusing and complex; there is no concept of self service in Data Boring. At the same time, the guys invested well in horizontal scaling, which is rare in the current market.

Filtering must be configured for each visualization separately. And there are great difficulties with cross-filtering, you will have to configure it via json config. Perhaps the technology will be improved this year; I would like a more user-friendly system. Those who, for example, like to use end-to-end filtering during transitions will not find it here. There are also difficulties with drill down; it can only be done along one dimension. If you have been spoiled by working with PBI, Tableau, Qlik, you will have to go through all the stages of adopting a new concept. In general, the MyzhSelfService experiment has not yet been a success.

Now about the calculations. The big disadvantages are that there are no variables, no aggr, and no measure wizard. Apparently method #1 is to pre-calculate everything at the loading stage, which in principle is sometimes even correct, especially on large data.

The pain of the market is the complex installation of software and a large number of open source components of non-local production under the hood. You have to sweat to unwrap it.

Colleagues need to focus more on community development like others are doing.