In Probing the roots of Samoyedic I note that already the clear fragmentation to separate languages demands a deeper age for Samoyedic than for the other Uralic subgroups. H.-W. Hatting asks in the comments a good argument-sharpening question, and writing my answer out in detail may fit better in a new blog post entirely.
His comment goes:
For that argumentation to work, the elimination of intermediate dialects in a continuum by replacement by other dominant dialects or by other languages would have to depend on the age of the language family, while it actually is caused by socio-economic factors that have nothing to do with that age. If Finnish would have been replaced starting from the Middle ages by (say) Swedish and Russian on a much larger scale than actually happened and at 1800 only a handful of dialects from extreme ends of the continuum would have survived, wouldn’t that look similar to Samoyedic? Or am I misunderstanding what you are saying here?
Very roughly, I’d say the issue is that a “fracturing” sociolinguistic phase does not yield itself very well to the accruement of further diversity across a language family. But we will also need to consider what are possible models for the fragmentation of Samoyedic at all.
At first pass, the biggest difference between this Finnish hypothetical and the situation in Samoyedic is of course that if an intrusive language has recently broken up a dialect continuum, we should be able to see the intruder. Yet Russian comes too late on the scene to be blamed for any of the boundaries between the six Samoyedic main groups. Take e.g. take Enets and Nganasan: they have “always” been right next to one another as far as the (meager) historical records go. Any possible transitional dialects must’ve been directly levelled out by Enets and Nganasan themselves, and/or before the Enets and Nganasan lineages arrived on the Taimyr peninsula. I don’t know of any other language either that could be plausibly blamed for splitting Selkup and the northern Samoyedic lineages off as their own groups. Sayan Turkic is the likely reason for leaving Mator and Kamassian in isolation… but as the two show no evidence of forming a common subgroup (their common isoglosses are all arealisms shared with Turkic!), even there this cannot be the only explanation.
This in mind, what looks to me like the fastest possible way to break Samoyedic apart into discrete lineages would be to assume that, after some initial dialect continuum development, some unknown language arrives on the scene and assimilates most of the core Samoyedic area. Then, once transitional varieties are on the way out, a handful of surviving marginal Samoyedic dialects re-gain momentum and expand again to become Nenets, Selkup, Kamassian–Koibal, etc.
The diversity we can observe today would mostly come about during the first and the third phase, and should sum up at least to the same two-ish millennia that less diverse families like Finnic and Samic show. But we also need to add some centuries for the whole fragmentation phase. Linguistic innovations would not be diffusing anymore between the late Common Samoyedic dialects at this time, slowing down the “native” component of language change. Superstrate influence can build up at a clip, but this should be roughly the same everywhere, not creating any substantial isoglosses.
The mystery intruder language does not need to go itself later extinct, at least not under the second expansion phase of Samoyedic. Something like Proto-Turkic could hence also work, with the end result being that none of the Samoyedic languages is anymore spoken anywhere near the original Proto-Samoyedic homeland. (Although in this particular case I’d expect to see much more Samoyedic influence ending up in the Turkicized varieties. Come to think of it, how much traces has the known Samoyedic substrate left in Sayan Turkic?)
Another option is that language boundaries within Samoyedic could be mostly “native”, based on secondary expansions like that of Tundra Nenets. This I think would fare even worse for a shallow dating, however.
I already expect linguistic expansions to be mostly serial rather than “explosive”. Cases like the European colonization of North America are rare (even that shows some spurts and lulls if you look closely). Since Finnish was raised as an example, it alone has gone through at least four major expansion phases:
- the initial phase, circa 0 CE, in establishing presence in the Kumo and Aura river valleys (likely at the expense of early Germanic speakers);
- the early middle ages’ agricultural expansion westward (probably mainly at the expense of early Sami varieties, tho I wouldn’t rule out some older “Lakelandic” substrate groups still being around too);
- the northwestern expansion around the Bay of Bothnia and into Lapland starting from the 15th century (again at the expense of Sami);
- and the Savonian expansion from the 17th century on, backed by the huuhtakaski method plus tax benefits from the Swedish crown (mopping up the last remaining Sami groups of southern Finland in the process).
Expansions always have some socioeconomic motivations too, which in almost all cases take time to come about or dissipate. (Eastern Finnish could only end up realistically marginalized and thoroughly splintered by 1800 if the main Savonian expansion never happened at all.)
But we would also need to explain how serial expansions could lead to fragmentation into separate languages entirely in the case of Samoyedic, unlike cases like Finnic. When a secondary expansion within a language family comes head-to-head with closely related dialects, the typical result is dialect mixing, not complete levelling. The Savonian expansion shows this well: towards its western edges, the northern Tavastian dialects of central Finland and the central Ostrobothnian dialects close to the coast (up as far as Oulu) end up picking up several Savonian / Eastern Finnish features, but also retaining several western features. The result is more diversity, not less.
Family-internal preliterate expansions  that eliminate dialect diversity can still happen: examples that come to mind are Slavic levelling out West Baltic, Dutch & Low German splintering Frisian, and indeed, Tundra Nenets levelling out Yurats. I get the impression however that this requires a more specific sociolinguistic setup. This would seem to only happen when a language variety expands so that it comes newly in contact with a relative that’s already different enough that they cannot re-converge back to a continuum anymore. Within a timespan of some two thousand years, most language families do not manage to pull this off even once! Assuming for Samoyedic some three-to-six of this event right in a row would be pretty unparsimonious. It seems inescapable to me that we would also in this scenario have to allow more time for Samoyedic to “stew”… so that e.g. a Proto-Selkup expansion can run into Para-Nenets or Para–Kamassian dialects that have already developed in rather different directions, and assimilate them without much trace of their earlier affiliation.
 By far the easiest way to level out dialect diversity is to roll in mass literacy + a standard literary language. This is not even remotely applicable to Samoyedic, of course.