About Us Products Baseline Extras Careers Contact us


Every Noesis service is integrated via API with ad-serving, DSP and SSP platforms. Currently, services are running on Open Adstream and AppNexus, but they are portable on every platform with a reasonable quality API.
Every service is extremely programmatic and founded on proprietary real time machine learning technologies.


Carousel maximizes the profitability of the available inventory while meeting advertisers’ goals. It already has a 4-year track record of success with publishers and agencies with managed inventory.
It delivers impressions from a certain page to campaigns promising the greatest profit, balancing this primary goal with the need to grant adequate service levels for less profitable advertisers (campaign completion, regular delivery).
It is much more complete than a standard purely auction-oriented system and ensures greater advertiser satisfaction. Moreover, campaign management costs are greatly reduced.
Note that Carousel optimization is not the same as that of AppNexus: it is smooth (delivering more or less) not threshold-driven (delivering or not at all, as AppNexus does).


SynAppSys is a programmatic bidder operating 24/7, choosing the best bid for each third-party inventory source, continuously adapting to changes in the marketplace environment.
It minimizes the campaign cost for inventory provisioning whilst at the same time ensuring budget completion, regularity of delivery and advertiser-specific goals such as a certain CTR or ECPA.
The key to SynAppSys’ effectiveness is the fine-tuning of bids at the inventory item level, driven by real time machine learning.
Note that SynAppSys optimization is not the same as that of AppNexus: it is smooth (delivering more or less) not threshold-driven (delivering or not at all, as AppNexus does).


Iceberg is a new kind of programmatic buyer. While “normal” buyers are essentially bidders, generating isolated bids, Iceberg builds up true buying strategies, a much richer concept.
In AppNexus, a line item describes the deal (e.g. revenue type is CPM) and the goals (e.g. the desired ECPA). It is up to the user to configure one or several campaigns implementing buying strategies supporting the deal and the goals. This is in itself an extremely challenging task, which becomes practically unfeasible if you want to adapt buying strategies in real time, let alone use data to optimize those strategies.
Indeed, Iceberg makes this AppNexus’ philosophy a concrete operation mode. It creates a number of campaigns (i.e. buying strategies) which implement the deal and the goals, and adapts those strategies 24/7 using real time feedback data.
While Iceberg seems to be strictly designed for inventory buyers, in fact it is perfectly suitable for publishers willing to offer their premium advertisers privileged access to premium inventory. Indeed, the first installation of the Iceberg service was with a big publisher.
Note that Iceberg optimization is not the same as that of AppNexus: it is smooth (delivering more or less) not threshold-driven (delivering or not at all, as AppNexus does).


Polyhedron is a programmatic “targetizer”: it reshapes and adapts a campaign targeting scheme 24/7 in order to best fit a campaign’s needs and features of the currently available inventory (both managed and third-party).
Usually a campaign manager chooses the campaign targets at the very beginning then checks delivery progress and performance level daily, possibly inserting and removing audience segments, geographic areas and so on. The target selection is intuitive, at best based on experience with past campaigns. Polyhedron algorithms perform target reshaping using campaign feedback data in a way which would be absolutely unfeasible for human users.
A groundbreaking feature of Polyhedron is multi-dimensionality. It simultaneously selects audience segments, pages, geos, devices, browsers, operating systems, day-parts. This amounts to evaluating thousands or even millions of multi-targeting options, a colossal task made manageable only by sophisticated proprietary algorithms. To the best of our knowledge, no other tool offers this kind of service.


GoodBuy is a dynamic selling engine, standing at the frontier between advertising and e-commerce. A single campaign includes many creatives, each one with its own economic value, for example, a flight with 5 alternative prices. The system delivers the right price to the right customer on the right page at the right time. What is optimized here is not a classic KPI like CTR or ECPA, but the true revenue.
The right price selection is only the beginning of the story: right product selection can be performed as well. Instead of one flight with five prices, you can have five products each one with its own price. Or you can have five different versions of the same product, selecting the right tailored configuration. The same holds for different bundling schemes of products.
In every case, the customer sees the “right” thing in a true economic sense. This is in no way similar to normal advertising campaign management, because it exploits not only optimization algorithms, but consumer predictive models, too.
With GoodBuy the distinction between advertising and selling almost vanishes, offering a great opportunity both for advertisers and e-commerce players, as well as for publishers and agencies suddenly able to offer a completely new kind of service while at the same time keeping their usual operations model.

Increase your portfolio
Not only do our products help you to do what you already do better.
With a little thought all kinds of new market offers come to mind.
Our current clients have already identified a whole range of new business opportunities which were impossible before we started working together.
What is success?
A difficult question.
Different for every context, for every client, both yours and ours.
Suffice to say that our clients and, in turn, their clients have answered this question in the affirmative time and again, whether their KPI was clickthrough, conversion rate, reach, right targeting or whatever.
Or something more qualitative.
Man vs. Machine?
Real time optimization is a must today. This is phenomenally complex.
There are some things that a machine can do which a person cannot. And vice versa.
Our background in fully automatized technology and AI machine learning means that our products do what is out of the question for a person, leaving the person to do what is out of the question for a machine.